Advances in technology have provided numerous innovations that make people's daily lives easier and more convenient

*1,015*
*220*
*37MB*

*English*
*Pages 578
[594]*
*Year 2019*

- Author / Uploaded
- Information Resources Management Association

*Table of contents : Title PageCopyright PageList of ContributorsTable of ContentsPrefaceSection 1: Chaotic and DNA-Based Cryptography Chapter 1: Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification in Pixel Domain and Chaotic Encryption Chapter 2: Chaotic Function Based ECG Encryption System Chapter 3: Biometric Image Security Using Chaos Algorithm Chapter 4: DNA Sequence Based Cryptographic Solution for Secure Image Transmission Chapter 5: IPHDBCMSection 2: Cryptographic Algorithms Chapter 6: Cryptographic Techniques Based on Bio-Inspired Systems Chapter 7: Modification of Traditional RSA into Symmetric-RSA Cryptosystems Chapter 8: Hybrid Approach of Modified AES Chapter 9: Cryptographic Algorithms for Next Generation Wireless Networks Security Chapter 10: Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network Chapter 11: Applicability of Cellular Automata in CryptanalysisSection 3: Encryption Keys and Homomorphic Encryption Chapter 12: A Novel Approach of Symmetric Key Cryptography using Genetic Algorithm Implemented on GPGPU Chapter 13: Provable Security for Public Key Cryptosystems Chapter 14: Towards Parameterized Shared Key for AVK Approach Chapter 15: Authentication of Smart Grid Chapter 16: Secure Speaker Recognition using BGN Cryptosystem with Prime Order Bilinear Group Chapter 17: A Pairing-based Homomorphic Encryption Scheme for Multi-User Settings Chapter 18: A Secure Cloud Storage using ECC-Based Homomorphic Encryption Chapter 19: Homomorphic Encryption as a Service for Outsourced Images in Mobile Cloud Computing EnvironmentSection 4: Steganography Chapter 20: Digital Image Steganography Chapter 21: Improved Secure Data Transfer Using Video Steganographic Technique Chapter 22: Secure Group Message Transfer Stegosystem Chapter 23: Implementation and Evaluation of Steganography Based Online Voting SystemSection 5: Visual Cryptography Chapter 24: Exploiting the Homomorphic Property of Visual Cryptography Chapter 25: Basic Visual Cryptography Using Braille Chapter 26: Threshold Secret Sharing Scheme for Compartmented Access Structures Chapter 27: An Improved Size Invariant (n, n) Extended Visual Cryptography Scheme Chapter 28: A Methodological Evaluation of Crypto-Watermarking System for Medical Images Chapter 29: Reversible Watermarking in Medical Image Using RDWT and Sub-Sample Chapter 30: Video Saliency Detection for Visual Cryptography-Based Watermarking Chapter 31: On the Pixel Expansion of Visual Cryptography Scheme Chapter 32: A Novel Pixel Merging-Based Lossless Recovery Algorithm for Basic Matrix VSS Chapter 33: A Contemplator on Topical Image Encryption MeasuresIndex*

Cryptography: Breakthroughs in Research and Practice Information Resources Management Association USA

Published in the United States of America by IGI Global Information Science Reference (an imprint of IGI Global) 701 E. Chocolate Avenue Hershey PA, USA 17033 Tel: 717-533-8845 Fax: 717-533-8661 E-mail: [email protected] Web site: http://www.igi-global.com Copyright © 2020 by IGI Global. All rights reserved. No part of this publication may be reproduced, stored or distributed in any form or by any means, electronic or mechanical, including photocopying, without written permission from the publisher. Product or company names used in this set are for identification purposes only. Inclusion of the names of the products or companies does not indicate a claim of ownership by IGI Global of the trademark or registered trademark. Library of Congress Cataloging-in-Publication Data Names: Information Resources Management Association, editor. Title: Cryptography : breakthroughs in research and practice / Information Resources Management Association, editor. Description: Hershey, PA : Information Science Reference, 2020. | Includes bibliographical references. | Summary: “”This book examines novel designs and recent developments in cryptographic security control procedures to improve the efficiency of existing security mechanisms that can help in securing sensors, devices, networks, communication, and data”-- Provided by publisher. Identifiers: LCCN 2019034135 (print) | LCCN 2019034136 (ebook) | ISBN 9781799817635 (h/c) | ISBN 9781799817642 (eISBN) Subjects: LCSH: Cryptography. | Data transmission systems--Security measures. | Computers--Access control. Classification: LCC TK5102.94 .C776 2020 (print) | LCC TK5102.94 (ebook) | DDC 005.8/24--dc23 LC record available at https://lccn.loc.gov/2019034135 LC ebook record available at https://lccn.loc.gov/2019034136 British Cataloguing in Publication Data A Cataloguing in Publication record for this book is available from the British Library. The views expressed in this book are those of the authors, but not necessarily of the publisher. For electronic access to this publication, please contact: [email protected].

List of Contributors

Abdelkader, Gafour / Djellali Liabes University, Algeria ...............................................................129 Agarwal, Sugandha / Amity University, India....................................................................................39 Agrawal, Nitesh Kumar / Indian School of Mines, India .................................................................449 Akojwar, Sudhir Gangadharrao / Government College of Engineering, India ..............................159 Alam, Naved / Jamia Hamdard, India ..............................................................................................180 Ali, Syed Taqi / National Institute of Technology Kurukshetra, India ...............................................214 Amine, Filali Mohamed / Djellali Liabes University, Algeria ..........................................................129 Anghelescu, Petre / University of Pitesti, Romania ............................................................................99 Annamalai, Murugan / Dr. Ambedkar Government Arts College, India ..........................................48 Ayyappan, Sonal / SCMS School of Engineering and Technology, India ........................................458 Babu, Anna / M. G. University, India ...............................................................................................458 Babu, Erukala Suresh / K.L. University, India ...................................................................................72 Banerjee, Anurag / M. S. Ramaiah Institute of Technology, India ..................................................193 Barari, Adrita / Defence Institute of Advanced Technology, India ..................................................498 Bhasin, Harsh / Jawahar Lal Nehru University, India .....................................................................180 Bhat, G.M. / University of Kashmir, India ............................................................................................1 Bhatia, Mahinder Pal Singh / Netaji Subhas Institute of Technology, India ....................................373 Bhatia, Manjot Kaur / University of Delhi, India ............................................................................373 Biswas, G. P. / Indian Institute of Technology (Indian School of Mines), India ........................120, 306 Dey, Nilanjan / Techno India College of Technology, India .................................................................1 Dhavale, Sunita V. / Defence Institute of Advanced Technology, India ............................................498 Dhawale, Chitra A. / P. R. Pote College of Engineering and Management, India............................332 Ding, Wanmeng / Hefei Electronic Engineering Institute, China ....................................................416 Driss, El Ouadghiri / MoulaySmail University, Morocco .................................................................316 Fathimal, P. Mohamed / Manonmaniam Sundaranar University, India ..........................................438 G. M., Siddesh / M. S. Ramaiah Institute of Technology, India........................................................193 Gao, Lin / Tianjin Chengjian University, China ...............................................................................480 Gao, Tiegang / Nankai University, China .........................................................................................480 Guo, Teng / University of International Relations, China ................................................................537 Gupta, Daya Sagar / Indian Institute of Technology (ISM) Dhanbad, India ....................................306 Haldar, Manas Kumar / Swinburne University of Technology, Malaysia ........................................391 Hassan, Naanani / Ben’msik University, Morocco ...........................................................................316 Hiriyannaiah, Srinidhi / M. S. Ramaiah Institute of Technology, India ..........................................193 Ibtihal, Mouhib / MoulaySmail University, Morocco ......................................................................316 Issac, Biju / Teesside University, UK ................................................................................................391 Jacob, Grasha / Rani Anna Government College, India ....................................................................48

Jambhekar, Naveen D. / S. S. S. K. R. Innani Mahavidyalaya, India ...............................................332 Jiao, Jian / University of International Relations, China .................................................................537 K. G., Srinivasa / M. S. Ramaiah Institute of Technology, India......................................................193 Khan, Zafar Jawed / RCERT Chandrapur, India .............................................................................159 Khare, Ayush / Indian School of Mines, India .................................................................................449 Kumari, R. Shantha Selva / Mepco Schlenk Engineering College, India ........................................277 Liu, Feng / Chinese Academy of Sciences, China .....................................................................428, 537 Liu, Hanlin / Hefei Electronic Engineering Institute, China ............................................................416 Liu, Lintao / Hefei Electronic Engineering Institute, China ............................................................416 Liu, Xin / Harbin Institute of Technology, China & Harbin University of Science and Technology, China ........................................................................................................................545 Lu, Yuliang / Hefei Electronic Engineering Institute, China ...........................................................416 Manna, G. C. / BSNL, India ..............................................................................................................142 Moh, Melody / San Jose State University, USA ................................................................................257 Mohit, Prerna / Indian Institute of Technology (Indian School of Mines), India .............................120 Mondal, Jayanta / KIIT University, India ........................................................................................556 Morappanavar, Anusha / M. S. Ramaiah Institute of Technology, India ........................................193 Muñoz, Melesio Calderón / Cupertino Electric, Inc., USA ..............................................................257 Muttoo, Sunil Kumar / University of Delhi, India............................................................................373 Nagaraju, C. / Y.V. University, India ..................................................................................................72 Nagaria, Deepak / Bundelkhand Institute of Engineering and Technology, India .............................39 Nidhyananthan, S. Selva / Mepco Schlenk Engineering College, India ...........................................277 Pal, Arup Kumar / Indian School of Mines, India ............................................................................449 Parah, Shabir A. / University of Kashmir, India ...................................................................................1 Prajapat, Shaligram / Maulana Azad National Institute of Technology, India & Devi Ahilya University, India ...........................................................................................................................239 Prasad, M. / Mepco Schlenk Engineering College, India .................................................................277 Prasad, M.H.M. Krishna / J.N.T. University, India ...........................................................................72 Rani, P. Arockia Jansi / Manonmaniam Sundaranar University, India ...........................................438 Reddy, V. Lokeswara / K.S.R.M College of Engineering, India .......................................................355 Rura, Lauretha / Swinburne University of Technology, Malaysia ...................................................391 Sang, Jianzhi / Harbin Institute of Technology, China .....................................................................545 Sawlikar, Alka Prasad / RCERT Chandrapur, India ........................................................................159 Sharma, Dixit / Guru Nanak Dev University, India ...........................................................................22 Sharma, Rahul / Indian School of Mines, India ...............................................................................449 Sheikh, Javaid A. / University of Kashmir, India ..................................................................................1 Singh, Butta / Guru Nanak Dev University, India ..............................................................................22 Singh, Manjit / Guru Nanak Dev University, India ............................................................................22 Singh, O.P. / Amity University, India ..................................................................................................39 Suryavanshi, Vishnu / GHRCE Nagpur, India ................................................................................142 Swain, Debabala / KIIT University, India ........................................................................................556 Thakur, Ramjeevan Singh / Maulana Azad National Institute of Technology, India ......................239 Wan, Song / Hefei Electronic Engineering Institute, China .............................................................416 Wang, Guangyu / Chinese Academy of Sciences, China & Auckland University of Technology, New Zealand .................................................................................................................................428 Wang, Shen / Harbin Institute of Technology, China .......................................................................545

Wang, Wen / Chinese Academy of Sciences, China .........................................................................537 Wei, Zhang / Engineering University of Chinese Armed Police Force, China .................................295 Yan, Wei Qi / Chinese Academy of Sciences, China & Auckland University of Technology, New Zealand .........................................................................................................................................428 Yan, Xuehu / Hefei Electronic Engineering Institute, China............................................................416 Zhang, Weizhe / Harbin Institute of Technology, China ..................................................................545 Zhao, Jie / Tianjin Chengjian University, China ..............................................................................480

Table of Contents

Preface.................................................................................................................................................... xi Section 1 Chaotic and DNA-Based Cryptography Chapter 1 Realization of a New Robust and Secure Watermarking Technique Using DC Coeﬃcient Modiﬁcation in Pixel Domain and Chaotic Encryption ..........................................................................1 Shabir A. Parah, University of Kashmir, India Javaid A. Sheikh, University of Kashmir, India Nilanjan Dey, Techno India College of Technology, India G.M. Bhat, University of Kashmir, India Chapter 2 Chaotic Function Based ECG Encryption System................................................................................22 Butta Singh, Guru Nanak Dev University, India Manjit Singh, Guru Nanak Dev University, India Dixit Sharma, Guru Nanak Dev University, India Chapter 3 Biometric Image Security Using Chaos Algorithm ..............................................................................39 Sugandha Agarwal, Amity University, India O.P. Singh, Amity University, India Deepak Nagaria, Bundelkhand Institute of Engineering and Technology, India Chapter 4 DNA Sequence Based Cryptographic Solution for Secure Image Transmission .................................48 Grasha Jacob, Rani Anna Government College, India Murugan Annamalai, Dr. Ambedkar Government Arts College, India Chapter 5 IPHDBCM: Inspired Pseudo Hybrid DNA Based Cryptographic Mechanism to Prevent Against Collabrative Black Hole Attack in Wireless Ad hoc Networks ............................................................72 Erukala Suresh Babu, K.L. University, India C. Nagaraju, Y.V. University, India M.H.M. Krishna Prasad, J.N.T. University, India

Section 2 Cryptographic Algorithms Chapter 6 Cryptographic Techniques Based on Bio-Inspired Systems .................................................................99 Petre Anghelescu, University of Pitesti, Romania Chapter 7 Modiﬁcation of Traditional RSA into Symmetric-RSA Cryptosystems.............................................120 Prerna Mohit, Indian Institute of Technology (Indian School of Mines), India G. P. Biswas, Indian Institute of Technology (Indian School of Mines), India Chapter 8 Hybrid Approach of Modiﬁed AES ....................................................................................................129 Filali Mohamed Amine, Djellali Liabes University, Algeria Gafour Abdelkader, Djellali Liabes University, Algeria Chapter 9 Cryptographic Algorithms for Next Generation Wireless Networks Security ...................................142 Vishnu Suryavanshi, GHRCE Nagpur, India G. C. Manna, BSNL, India Chapter 10 Eﬃcient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network ...159 Alka Prasad Sawlikar, RCERT Chandrapur, India Zafar Jawed Khan, RCERT Chandrapur, India Sudhir Gangadharrao Akojwar, Government College of Engineering, India Chapter 11 Applicability of Cellular Automata in Cryptanalysis .........................................................................180 Harsh Bhasin, Jawahar Lal Nehru University, India Naved Alam, Jamia Hamdard, India Section 3 Encryption Keys and Homomorphic Encryption Chapter 12 A Novel Approach of Symmetric Key Cryptography using Genetic Algorithm Implemented on GPGPU ...............................................................................................................................................193 Srinivasa K. G., M. S. Ramaiah Institute of Technology, India Siddesh G. M., M. S. Ramaiah Institute of Technology, India Srinidhi Hiriyannaiah, M. S. Ramaiah Institute of Technology, India Anusha Morappanavar, M. S. Ramaiah Institute of Technology, India Anurag Banerjee, M. S. Ramaiah Institute of Technology, India

Chapter 13 Provable Security for Public Key Cryptosystems: How to Prove that the Cryptosystem is Secure ...214 Syed Taqi Ali, National Institute of Technology Kurukshetra, India Chapter 14 Towards Parameterized Shared Key for AVK Approach ....................................................................239 Shaligram Prajapat, Maulana Azad National Institute of Technology, India & Devi Ahilya University, India Ramjeevan Singh Thakur, Maulana Azad National Institute of Technology, India Chapter 15 Authentication of Smart Grid: The Case for Using Merkle Trees ......................................................257 Melesio Calderón Muñoz, Cupertino Electric, Inc., USA Melody Moh, San Jose State University, USA Chapter 16 Secure Speaker Recognition using BGN Cryptosystem with Prime Order Bilinear Group ...............277 S. Selva Nidhyananthan, Mepco Schlenk Engineering College, India M. Prasad, Mepco Schlenk Engineering College, India R. Shantha Selva Kumari, Mepco Schlenk Engineering College, India Chapter 17 A Pairing-based Homomorphic Encryption Scheme for Multi-User Settings....................................295 Zhang Wei, Engineering University of Chinese Armed Police Force, China Chapter 18 A Secure Cloud Storage using ECC-Based Homomorphic Encryption .............................................306 Daya Sagar Gupta, Indian Institute of Technology (ISM) Dhanbad, India G. P. Biswas, Indian Institute of Technology (ISM) Dhanbad, India Chapter 19 Homomorphic Encryption as a Service for Outsourced Images in Mobile Cloud Computing Environment ........................................................................................................................................316 Mouhib Ibtihal, MoulaySmail University, Morocco El Ouadghiri Driss, MoulaySmail University, Morocco Naanani Hassan, Ben’msik University, Morocco Section 4 Steganography Chapter 20 Digital Image Steganography: Survey, Analysis, and Application .....................................................332 Chitra A. Dhawale, P. R. Pote College of Engineering and Management, India Naveen D. Jambhekar, S. S. S. K. R. Innani Mahavidyalaya, India

Chapter 21 Improved Secure Data Transfer Using Video Steganographic Technique ..........................................355 V. Lokeswara Reddy, K.S.R.M College of Engineering, India Chapter 22 Secure Group Message Transfer Stegosystem ....................................................................................373 Mahinder Pal Singh Bhatia, Netaji Subhas Institute of Technology, India Manjot Kaur Bhatia, University of Delhi, India Sunil Kumar Muttoo, University of Delhi, India Chapter 23 Implementation and Evaluation of Steganography Based Online Voting System ..............................391 Lauretha Rura, Swinburne University of Technology, Malaysia Biju Issac, Teesside University, UK Manas Kumar Haldar, Swinburne University of Technology, Malaysia Section 5 Visual Cryptography Chapter 24 Exploiting the Homomorphic Property of Visual Cryptography .......................................................416 Xuehu Yan, Hefei Electronic Engineering Institute, China Yuliang Lu, Hefei Electronic Engineering Institute, China Lintao Liu, Hefei Electronic Engineering Institute, China Song Wan, Hefei Electronic Engineering Institute, China Wanmeng Ding, Hefei Electronic Engineering Institute, China Hanlin Liu, Hefei Electronic Engineering Institute, China Chapter 25 Basic Visual Cryptography Using Braille...........................................................................................428 Guangyu Wang, Chinese Academy of Sciences, China & Auckland University of Technology, New Zealand Feng Liu, Chinese Academy of Sciences, China Wei Qi Yan, Chinese Academy of Sciences, China & Auckland University of Technology, New Zealand Chapter 26 Threshold Secret Sharing Scheme for Compartmented Access Structures ........................................438 P. Mohamed Fathimal, Manonmaniam Sundaranar University, India P. Arockia Jansi Rani, Manonmaniam Sundaranar University, India Chapter 27 An Improved Size Invariant (n, n) Extended Visual Cryptography Scheme ......................................449 Rahul Sharma, Indian School of Mines, India Nitesh Kumar Agrawal, Indian School of Mines, India Ayush Khare, Indian School of Mines, India Arup Kumar Pal, Indian School of Mines, India

Chapter 28 A Methodological Evaluation of Crypto-Watermarking System for Medical Images........................458 Anna Babu, M. G. University, India Sonal Ayyappan, SCMS School of Engineering and Technology, India Chapter 29 Reversible Watermarking in Medical Image Using RDWT and Sub-Sample ....................................480 Lin Gao, Tianjin Chengjian University, China Tiegang Gao, Nankai University, China Jie Zhao, Tianjin Chengjian University, China Chapter 30 Video Saliency Detection for Visual Cryptography-Based Watermarking ........................................498 Adrita Barari, Defence Institute of Advanced Technology, India Sunita V. Dhavale, Defence Institute of Advanced Technology, India Chapter 31 On the Pixel Expansion of Visual Cryptography Scheme ..................................................................537 Teng Guo, University of International Relations, China Jian Jiao, University of International Relations, China Feng Liu, Chinese Academy of Sciences, China Wen Wang, Chinese Academy of Sciences, China Chapter 32 A Novel Pixel Merging-Based Lossless Recovery Algorithm for Basic Matrix VSS ........................545 Xin Liu, Harbin Institute of Technology, China & Harbin University of Science and Technology, China Shen Wang, Harbin Institute of Technology, China Jianzhi Sang, Harbin Institute of Technology, China Weizhe Zhang, Harbin Institute of Technology, China Chapter 33 A Contemplator on Topical Image Encryption Measures ..................................................................556 Jayanta Mondal, KIIT University, India Debabala Swain, KIIT University, India Index ...................................................................................................................................................574

xi

Preface

Technology in today’s world has developed rapidly with the emergence of smart computation and artificial intelligence. These devices continue to integrate themselves into various fields of research and practice, making operations and tasks easier to accomplish. Despite the vast benefits technology has to offer, the swift development has created an array of vulnerabilities and unlocked a significant amount of security and privacy risks. The protection of valuable data and information has become a key topic of research because of this. Cryptography is the practice of writing codes and developing secure communication techniques. The art of cryptography has emerged as a prevalent area of study as many practitioners and professionals are relying on this technology to keep their information safe from cyber attacks and hackers. Within this up-and-coming subject lies a diverse amount of methods and approaches that can be specifically applied to countless professions. Because of cryptography’s popularity, there are new studies and research findings being published at a rapid pace. It remains a challenge for the general public to keep up with the constant discoveries being made in this field. Researchers, universities, and professionals need a compilation of information that stays up to date on the latest findings and studies regarding cryptography and provides them with current advancements and future trends in this industry. The everchanging landscape surrounding the diverse applications of different scientific areas can make it very challenging to stay on the forefront of innovative research trends. That is why IGI Global is pleased to offer this one-volume comprehensive reference that will empower academicians, graduate students, engineers, IT specialists, software engineers, security analysts, industry professionals, and researchers with a stronger understanding of cryptography. This compilation is designed to act as a single reference source on conceptual, methodological, and technical aspects and will provide insight into emerging topics including but not limited to cybersecurity, encryption, information security, intrusion detection systems, authentication, and threat detection. The chapters within this publication are sure to provide readers the tools necessary for further research and discovery in their respective industries and/or fields. Cryptography: Breakthroughs in Research and Practice is organized into five sections that provide comprehensive coverage of important topics. The sections are: 1. 2. 3. 4. 5.

Chaotic and DNA-Based Cryptography; Cryptographic Algorithms; Encryption Keys and Homomorphic Encryption; Steganography; and Visual Cryptography.

Preface

The following paragraphs provide a summary of what to expect from this invaluable reference source: Section 1, “Chaotic and DNA-Based Cryptography,” opens this extensive reference source by highlighting the latest trends in DNA-based cryptography techniques and explores the application of the mathematical chaos theory to the practice of cryptography. The first chapter in this section, “Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification in Pixel Domain and Chaotic Encryption,” by Profs. Shabir Parah and Javaid Sheikh from the University of Kashmir, India; Prof. Nilanjan Dey from Techno India College of Technology, India; and Prof. G.M. Bhat from the University of Kashmir, India, explores the security of digital watermarking using new encryption methods and embedded systems. Another featured chapter, “Chaotic Function Based ECG Encryption System,” by Profs. Butta Singh, Manjit Singh, and Dixit Sharma from Guru Nanak Dev University, India, studies specific encryption structures for remote healthcare monitoring systems and protecting the integrity of medical data. The following chapter, “Biometric Image Security Using Chaos Algorithm,” by Profs. Sugandha Agarwal and O.P. Singh from Amity University, India and Prof. Deepak Nagaria from Bundelkhand Institute of Technology, India, discusses experimental analyses that provide an efficient and secure algorithm for real-time image encryption and transmission. One of the closing chapters, “DNA Sequence Based Cryptographic Solution for Secure Image Transmission,” by Prof. Grasha Jacob from Rani Anna Government College, India and Prof. Murugan Annamalai from Dr. Ambedkar Government Arts College, India, features research on an encryption scheme based on DNA sequences that enable secure transmission of images across the internet. The final chapter in this section is “IPHDBCM: Inspired Pseudo Hybrid DNA Based Cryptographic Mechanisms to Prevent against Collaborative Black Hole Attack in Wireless Ad hoc Networks,” authored by Prof. Erukala Suresh Babu from K.L. University, India; Prof. C. Nagaraju from Y.V. University, India; and Prof. M.H.M. Krishna Prasad from J.N.T. University, India. This chapter addresses the detection and defense of blackhole attacks using hybrid DNA-based cryptography (HDC) mechanisms. Section 2, “Cryptographic Algorithms,” includes chapters on emerging innovations and applications for optimized algorithms within cryptography. The initial chapter in this section, “Cryptographic Techniques Based on Bio-Inspired Systems,” by Prof. Petre Anghelescu from the University of Pitesti, Romania, provides an alternative to conventional security methods using bio-inspired techniques based on the cellular automata (CAs) and programmable cellular automata (PCAs) theory. The following chapter is “Modification of Traditional RSA into Symmetric-RSA Cryptosystems,” authored by Profs. Prerna Mohit and G.P. Biswas from the Indian Institute of Technology, India. This chapter addresses the modification of RSA cryptography using extensive applications including digital signature and encryption. Another chapter in this section, “Hybrid Approach of Modified AES,” by Profs. Filali Mohamed Amine and Gafour Abdelkader from Djellali Liabes University, Algeria, discusses the modification and transformation of advanced encryption standard algorithms using images as input data. The chapter that follows this, “Cryptographic Algorithms for Next Generation Wireless Networks Security,” is written by Prof. Vishnu Suryavanshi from GHRCE Nagpur, India and Prof. G.C. Manna from BSNL, India. This chapter explores the encoding of information and communication security in next generation networks by using quantum cryptographic algorithms. Another chapter found within this section, “Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network,” by Profs. Alka Prasad Sawlikar and Zafar Jawed Khan from RCERT Chandrapur, India and Prof. Sudhir Gangadharrao Akojwar from the Government College of Engineering, India, examines the application of conversion and encryption techniques to secure and compress transmitted data. The closing chapter, “Applicability of

xii

Preface

Cellular Automata in Cryptanalysis,” authored by Prof. Harsh Bhasin from Jawar Lal Nehru University, India and Prof. Naved Alam from Jamia Hamdard, India, explores the applicability of soft computing techniques within cryptanalysis and discusses the future scope of similar methods. Section 3, “Encryption Keys and Homomorphic Encryption,” presents coverage on novel strategies within the field of encryption and key technology. The opening chapter of this section, “Provable Security for Public Key Cryptosystems: How to Prove that the Cryptosystem is Secure,” by Prof. Syed Taqi Ali from National Institute of Technology Kurukshetra, India, reviews the security proofs of well known public key cryptosystems as well as the multiple approaches for structuring these proofs. This chapter is followed by “Towards Parameterized Shared Key for AVK Approach,” which is written by Prof. Shaligram Prajapat and Prof. Ramjeevan Singh Thakur from the Maulana Azad National Institute of Technology, India. This chapter examines the implementation of the automatic variable key (AVK) approach within symmetric key cryptosystems from various users’ perspectives. The following chapter, “Authentication of Smart Grid: The Case for Using Merkle Trees,” by Prof. Melesio Calderón Muñoz from Cupertino Electric, Inc., USA and Prof. Melody Moh from San Jose State University, USA, argues for the use of Merkle trees as opposed to public key encryption for the authentication and security of devices within smart grids. Another noteworthy chapter found in this section is “Secure Speaker Recognition using BGN Cryptosystem with Prime Order Bilinear Group,” by Profs. S. Selva Nidhyananthan, Prasad M., and Shantha Kumari R. from Mepco Schlenk Engineering College, India. This chapter examines the efficiency of secure speaker recognition frameworks by implementing BGN Cryptosystems. One of the following chapters, “A Pairing-based Homomorphic Encryption Scheme for Multi-User Settings,” by Prof. Zhang Wei from the Engineering University of Chinese Armed Police Force, China, provides detailed security analysis of a pairing-based multi-user homomorphic encryption scheme by adopting the idea of proxy re-encryption and focusing on the compatibility of computation. Also included in this section is the chapter “A Secure Cloud Storage using ECC-Based Homomorphic Encryption,” by Profs. Daya Sagar Gupta and G.P. Biswas from the Indian Institute of Technology Dhanbad, India. This chapter presents a new homomorphic public-key encryption scheme based on elliptic curve cryptography, which allows public computation on encrypted data stored on a cloud. The closing chapter of this section, “Homomorphic Encryption as a Service for Outsourced Images in Mobile Cloud Computing Environment,” written by Profs. Mouhib Ibtihal and El Ouadghiri Driss from MoulaySmail University, Morocco and Prof. Naanani Hassan from Ben’msik University, Morocco, proposes a two cloud approach, a private cloud dedicated to encryption and a public cloud dedicated to storage, for the security and protection of outsourced images in mobile cloud computing atmospheres. Section 4, “Steganography,” discusses coverage and research perspectives on the latest developments and implementations of steganographic methods. In the foremost chapter of this section, “Digital Image Steganography: Survey, Analysis, and Application,” by Prof. Chitra A. Dhawale from P. R. Pote College of Engineering and Management, India and Prof. Naveen D. Jambhekar from S. S. S. K. R. Innani Mahavidyalaya, India, the authors discuss the basics of digital image steganographic techniques as well as types of images used, performance analysis of various steganographic algorithms used for attacks, and current applications. Following this chapter is “Improved Secure Data Transfer Using Video Steganographic Technique,” authored by Prof. V. Lokeswara Reddy from K.S.R.M College of Engineering, India. This chapter explores the modern development of video steganography by using a double layered security technique. Another chapter presented in this section, “Secure Group Message Transfer Stegosystem,” by Prof. Mahinder Pal Singh Bhatia from the Netaji Subhas Institute of Technology, India and

xiii

Preface

Profs. Manjot Kaur Bhatia and Sunil Kumar Muttoo from the University of Delhi, India, proposes a new secure message broadcasting system to hide the messages so attackers cannot sense the existence of messages, using steganography and image encryption. The closing chapter within this section, “Implementation and Evaluation of Steganography Based Online Voting System,” by Prof. Lauretha Rura from Swinburne University of Technology, Malaysia; Prof. Biju Issac from Teesside University, UK; and Prof. Manas Kumar Haldar from Swinburne University of Technology, Malaysia, studies novel approaches to online voting by combining visual cryptography with image steganography to improve system security without damaging usability and performance. Section 5, “Visual Cryptography,” highlights the latest research findings pertaining to cryptographic techniques that allow visual information to be encrypted. The opening chapter of this section, “Exploiting the Homomorphic Property of Visual Cryptography,” by Profs. Xuehu Yan, Yuliang Lu, Lintao Liu, Song Wan, Wanmeng Ding, and Hanlin Liu from the Hefei Electronic Engineering Institute, China, demonstrates the effectiveness and security of homomorphic visual cryptographic scheme (HVCS) with theoretical analysis and simulation results. This chapter is followed by “Basic Visual Cryptography Using Braille,” which is authored by Profs. Guangyu Wang, Feng Liu, and Wei Qi Yan from the Chinese Academy of Sciences, China. In this chapter, the authors conduct an experiment in which they embed Braille into visual cryptography (VC) methods due to the similarities in both approaches, with the intent to enhance the security of VC shares. Another chapter featured in this section, “Threshold Secret Sharing Scheme for Compartmented Access Structures,” written by Profs. P. Mohamed Fathimal and P. Arockia Jansi Rani from Manonmaniam Sundaranar University, India, discusses a new method of secret image sharing that eliminates the danger of image regeneration and misuse, as well as offers better visual quality of recovered images using a compartmented scheme. Another early chapter presented within this section, “An Improved Size Invariant (n, n) Extended Visual Cryptography Scheme,” by Profs. Rahul Sharma, Nitesh Kumar Agrawal, Ayush Khare, and Arup Kumar Pal from the Indian School of Mines, India, examines a protracted cryptography method using a distinct number of image shares that is able to generate a visually secret message without extending its size while also enhancing the visual quality. Also contained within this section is the chapter “A Methodological Evaluation of Crypto-Watermarking System for Medical Images,” written by Prof. Anna Babu from M.G. University, India and Prof. Sonal Ayyappan from SCMS School of Engineering and Technology, India. This chapter presents the basic aspects of crypto-watermarking techniques and their application within the secure transmission of medical images. Following this chapter is “Reversible Watermarking in Medical Image Using RDWT and Sub-Sample,” by Profs. Lin Gao and Tiegang Gao from Nankai University, China and Prof. Jie Zhao from Tianjin Chengjian University, China, which explores a reversible medical image watermarking scheme that uses redundant discrete wavelet transform (RDWT) and sup-sample in order to meet the high demand of perceptional quality as well as providing enhancement on the embedding capacity. Another noteworthy chapter, “Video Saliency Detection for Visual Cryptography-Based Watermarking,” by Profs. Adrita Barari and Sunita V. Dhavale from the Defence Institute of Advanced Technology, India, reviews the application of visual cryptography techniques in non-intrusive video watermarking with an emphasis on saliency feature extraction in videos. One of the closing chapters in the section, “On the Pixel Expansion of Visual Cryptography Scheme,” is written by Profs. Teng Guo and Jian Jiao from the School of Information Science and Technology, China and Profs. Feng Liu and Wen Wang from the Chinese Academy of Sciences, China. This chapter discusses the pixel expansion of visual cryptography methods within various kinds of graph access structures as well as graph decomposition techniques derived from these schemes. Another chapter featured in the culmination of this section, “A xiv

Preface

Novel Pixel Merging-Based Lossless Recovery Algorithm for Basic Matrix VSS,” authored by Profs. Xin Liu, Shen Wang, Jianzhi Sang, and Weizhe Zhang from Harbin Institute of Technology, China, analyzes a lossless recovery algorithm for visual secret share (VSS) that implements exclusive XOR operation and merging pixel expansion. The final chapter, “A Contemplator on Topical Image Encryption Measures,” by Profs. Jayanta Mondal and Debabala Swain from KIIT University, India, illustrates a survey on image encryption in various domains providing an introduction to cryptography as well as developing a review of sundry techniques. Although the primary organization of the contents in this work is based on its five sections, offering a progression of coverage of the important concepts, methodologies, technologies, applications, social issues, and emerging trends, the reader can also identify specific contents by utilizing the extensive indexing system listed at the end.

xv

Section 1

Chaotic and DNA-Based Cryptography

1

Chapter 1

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification in Pixel Domain and Chaotic Encryption Shabir A. Parah University of Kashmir, India

Nilanjan Dey Techno India College of Technology, India

Javaid A. Sheikh University of Kashmir, India

G.M. Bhat University of Kashmir, India

ABSTRACT The proliferation of information and communication technology has made exchange of information easier than ever. Security, Duplication and manipulation of information in such a scenario has become a major challenge to the research community round the globe. Digital watermarking has been found to be a potent tool to deal with such issues. A secure and robust image watermarking scheme based on DC coefficient modification in pixel domain and chaotic encryption has been presented in this paper. The cover image has been divided into 8×8 sub-blocks and instead of computing DC coefficient using Discrete Cosine Transform (DCTI, the authors compute DC coefficient of each block in spatial domain. Watermark bits are embedded by modifying DC coefficients of various blocks in spatial domain. The quantum of change to be brought in various pixels of a block for embedding watermark bit depends upon DC coefficient of respective blocks, nature of watermark bit (0 or 1) to be embedded and the adjustment factor. The security of embedded watermark has been taken care of by using chaotic encryption. Experimental investigations show that besides being highly secure the proposed technique is robust to both signal processing and geometric attacks. Further, the proposed scheme is computationally efficient as DC coefficient which holds the watermark information has been computed in pixel domain instead of using DCT on an image block.

DOI: 10.4018/978-1-7998-1763-5.ch001

Copyright © 2020, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification

1. INTRODUCTION The advancement in communication and networked and multimedia technologies and exponential rise in the users of internet world-wide has resulted in reproduction and distribution of multimedia content likeaudio, images and videos easier. In such a protection of multimedia content has become one of the prominent issues. Various encryption techniques are being used to encrypt the multimedia information before actual data transmission to avert various security and Intellectual Property Right (IPR) problems. However, the disguised look of the scrambled data makes the attacker more suspicious and hence the chances of a malicious attack from the adversary get increased. Given the significance of the problem some serious work needs to be done in order to ensure security and maintain the easy availability of multimedia content. In recent years digital watermarking has received most attention for security and protect multimedia data (Cox et al, 1998; Djurovic et al, 2001; Parah et al, 2014a). A digital watermark is a special data such as logo, imperceptibly embedded in multimedia content like an image etc. to prove its ownership. Since images are one of the prominent members of multimedia content, most of the developed watermark schemes reported till date use images as cover media (Ghouti, et al, 2006). Depending upon visibility of watermark, watermarking schemes are classified into two classes viz.; visible and invisible techniques. Most generally invisible (imperceptible) watermarking is used for copyright protection. Ina typical imperceptible watermarking technique,the watermark or special information datais embedded inside a cover image in such a way that it is imperceptible. Thus, it does not catch the attention of human visual system and protectsthe cover image from common signal processing and geometric attacks. The aim is to create a watermarked image that looks precisely same to a human eye but ensures ownership claim whenever necessary. Digital watermarking has been successfully validated to be very suitable in identifying thesource; creator, owner and distributor of a digital multimedia object (Shih, 2008). Digital image watermarking techniques are classified into various classes depending on various laid criteria. One of the prominent classifications is based on the domain of embedding the watermark. Based on this criteria watermarking is classified into spatial and transform domain (Shabir et al, 2013c; Parah et al, 2015d). Spatial domain watermarking techniques are the earliest and simplest. In spatial domain watermarking the watermark is embedded in some of the selected pixels (or all) of a cover image (Shabir et al, 2012b; Shabir et al, 2013c; Shabir et al, 2012c; Parah et al, 2015c). On the other hand, transform domain watermarking techniques involve modification of the transformed coefficients of the cover image. These transform domain watermarking makes use of various image transforms like Fourier Transform (FT) (Cintra et al, 2009; Liu and Zaho, 2010), Discrete Wavelet Transform (DWT) (Ghuti et al, 2006; Lu et al, 2012; Tsai, 2011; Wang et al, 2007), Singular Value Decomposition (SVD) (Djurovic et al, 2001; Lai and Tsai, 2010; Liu and Tan, 2002; Chen et al, 2013), Fractional Fourier Transform (FFT) (Bhatnagar and Ramman, 2011) and Contorlet transform. Pixel domain (spatial domain) watermarking schemes have least computational overhead; however they are fragile to various image processing and geometric attacks (Shabir et al, 2014a; Shabir et al, 2013a; Shabir et al, 2012a; Shabir et al, 2015; Shabir et al, 2014c). Transform domain methods on the other hand, are robust as compared to spatialdomain techniques. It is due to the underlying fact that when the inverse transformation is applied to a watermarked image, the watermark is irregularly distributed over the whole image. Thus it is very difficult for an attacker to extract or even modify the watermark. This paper presents a very intresting approach to watermarking, wherein we have successfully shown that a robust watermarking system can be implemented in spatial domain by embedding the watermark in DC component of Discrete Cosine Transfer (DCT) coefficients. Rest of the paper has been organised as follows. An extensive survey of 2

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification

literature has been presented in section 2. Mathematical preliminaries have been discussed in section 3. Proposed system has been discussed in detail in section 4. Section 5 presents experimental results and discussions. The paper concludes in section 6.

2. RELATED WORK A lot of research has been reported in the field of digital watermarking. Parah et al, 2016 have proposed a robust and blind watermarking technique in DCT domain using inter-block coefficient differencing. The scheme utilises the difference between two DCT coefficients of the adjacent blocks to embed the watermark information. The scheme has been shown to be robust to various image processing andgeometric attacks. The authors show that besides singular attacks the scheme is robust to hybrid attacks as well. An adaptive blind image watermarking using edge pixel concentration has been reported in (Fazlali et al, 2016). A two-level Contourlet transform on the cover image has been used. Thefirst level approximate image is partitioned into blocks and edge information is extracted. This is followed by application of DCT Transform at the block level. Improved robustness has been guaranteed by embedding the watermark redundantly and using majority vote during extraction.A novel watermarking system for copyright protection and verification has been proposed in (Bhatnagr and Wu, 2015). The cover image has been segmented into non-overlapping blocks utilizing the means of space filling curve amount of DCT energy in the blocks. Multiple watermarks have been embedded in the cover image. Singular Value Decomposition (SVD) has been used to embed watermarks into the image. The scheme besides yielding good perceptual quality watermark images has been shown to be robust to various image processing attacks. A frequency domain adaptive digital image watermarking scheme has been reported in (Karla et al, 2015). The scheme has been designed for color images. Prior to watermark embedding Hamming codes are added to the intensity component of color image. Dual encryption has been used on watermark to ensure its security. The scheme makes use of DWT, DCT and Arnold transform and has been reported to be robust to various signal processing attacks. A blind watermarking technique for images based on chaotic mixtures has been reported in (Niansheng et al, 2015). Logistic maps have been firstly used to scramble the watermark. DWT is used on cover image and watermark is embedded approximation coefficients. The authors proposed a technique displayingvery good properties with regard to imperceptibility, robustness and security.A blind watermarking scheme where middle frequency bands of the Fractional Fourier Transform (FrFT) of an image are used to embed watermark has been reported in (Lang and Zang, 2014). The scheme is robust to manyimageprocessing operations. But the proposed system has less robustness to JPEG compression.A secure and robust watermarking technique for gray scale images is presented in (Guo et al, 2015). Both cover image and logo are encrypted before embedding. However, the schemeperformance to JPEG compression waspoor.A DCT based watermarking technique has been reported in (Ma et al, 2012). Arnold transform has been used to reduce the correlation between pixels and enhance security. A watermarking technique based on the concept of mathematical remainder is reported in (Lin et al, 2010).The scheme was robust against various image processing attacks has lesser imperceptibility. A DCT based robust watermarking technique has been reported in (Das et al, 2014). The cover image was divided into 8 ×8 pixel blocks followed by computation of DCT of each block. The scheme has been reported to be robust to various image processing and geometric attacks.

3

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification

All the above discussions reveal that watermarking is usually carried out in spatial domain or frequency domain with both the domains. The main advantage of former is lesser computational complexity and the later robustness to various attacks. In this paper we try to explore a watermarking system which has lesser computational complexity like spatial domain but offer robustness to various attacks as in transform domain. We have successfully shown that a robust watermarking system can be implemented in spatial domain by embedding the watermark in DC component of DCT. It is however important to mention that DCT has not been used for computation of DC component of a block, but it has been generated in spatial domain.It is in place to mention that (Shih and Wu, 2003) have already put forth watermarking scheme which uses both spatial domain and frequency domain concepts However, unlike the proposed scheme the authors make use of actual DCT transform to compute various coefficients. Consequently, a new blind watermarking schemegrey scale images has been proposed in this paper. The scheme makes use of arithmetic mean of a selected block to compute its corresponding DC coefficient (without involving use of DCT, DC coefficient is computed). The watermark is embedded by modifying pixel values in accordance with watermark informationand the quantization step. The proposed system not only evades the errors resulting from the spectral transformation but also keepsthe distribution feature of frequency coefficients. As such the proposed scheme is computationally simpler.

3. MATHEMATICAL PRELIMINARIES 3.1. Computation DC Coefficients in Spatial Domain Discrete Cosine Transform (DCT) is used for domain transformation of a given set of real numbers. The transform kernel of DCT is cosine function. DCT has been used widely in various multimedia watermarking applications due to the fact that famous compression standard JPEG also makes use of it. For the transformation of images from spatial domain to spectral domain 2-D DCT transform is used. Conversely for restoration of an image from its frequency components the inverse 2-D DCT is used. For a P × Q image g (x, y) (x = 0, 1, 2,. . ., P-1, y = 0, 1, 2,. . ., Q-1),its 2-D DCT is given as follows: P −1 Q −1

C (u, v ) = αu αv ∑∑g (x , y ) cos

π ( 2x + 1) u 2P

x =0 y =0

cos

π ( 2y + 1) v 2Q

(1)

where P and Qrepresent the rows and columnsof g(x, y), u and v are the horizontal and vertical frequency components (u = 0, 1, 2,. . ., P-1, v = 0, 1, 2,. . .,Q-1), C(u, v) is the DCT coefficient of image g(x, y). αu =

4

1 P

for u = 0

2 for 1 ≤ u < P − 1 P

(2)

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification

αv =

1 Q

for v = 0

2 for 1 ≤ v < Q − 1 Q

(3)

The inverse DCT is given by: P −1 Q −1

g (x , y ) = ∑∑αu αv C (u,v ) cos

π ( 2x + 1) u

u =0 v =0

2P

cos

π ( 2y + 1) v 2Q

(4)

Expanding the above equation we get: g (x , y ) = α0α0C (0, 0) + α0α1C (0, 1) cos +...... + αP αQ C (P,Q ) cos

π ( 2y + 1) 2Q

P π ( 2x + 1) 2P

cos

+ α0α2C (0, 2) Q π ( 2y + 1)

(5)

2Q

It is evident from the above expanded expression that spatial domain image can be obtained from various constituent frequency domain coefficients. The DC coefficient in the DCT domain can be easily found using equation (1) and is given as: C (0, 0) =

1 PQ

P −1 Q −1

∑∑g (x, y )

(6)

x =0 y =0

It is evident from Equation (6) that DC coefficient C(0,0) is simply averaged sum of all pixel values of g(x, y) in the pixel domain The basic procedure of adding a watermark in DCT domain involves addition of watermark information to various DCT coefficients, followed by usage of inverse DCT to obtain watermarked image. It is a proven fact that the energy of signal added to DC coefficient does not suffer any loss after the application of inverse DCT. In order to elaborate the results, the authors can refer to (Qingtang et al, 2013). The outline of the whole process is that embedding watermark into the DC coefficient in DCT domain can be easily replaced in the pixel domain.

3.2. DC Component Modification We have already shown that the DC coefficient can be obtained by using arithmetic average of a given image block in spatial domain. Further,watermark embedding the DC component of DCT domain can be achieved by adjusting the value of pixel in the spatial domain appropriately. It is however pertinent to mention that the modified total of all the picture elements in pixel domain must equal thealtered value of DC coefficient in transform domain. It is significant to find the updating value of every pixel in the pixel domain in accordance with changed value of DC component in the transform domain.

5

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification

From Equation (4), the inverse DCT can be rewritten as: g (x , y ) =

1 PQ

C (0, 0) + g ˜(x , y )

(7)

where g~(x,y) represents the reconstructed image from AC components. Assume that cover image is represented by number of non-overlapping blocks represented as: M N g (x , y ) = gi, j (m, n ), 0 ≤i < , 0 ≤ j < , 0 ≤m, n 0 && x(y)=0.5 ew(y)=data(y); %% ‘ew’ stores the final encrypted watermark end end

4.2. Watermark Embedding The applied embed the watermark in a cover image are as follows: 1) read the cover image and divide it into 8 × 8 non-overlapping pixel blocks, 2) encrypt the watermark (as per section 3.1), 3) directly compute DC coefficient Cij(0,0) of each block according to Equation (6), and iv) asper encrypted watermark bit value ew(i, j) alteration factors magnitudes AF1and AF2 as shown below are chosen: AF 1 = 0.5β if ew (i, j ) = 1, AF 2 = −1.5β

(17)

AF 1 = −0.5β if ew (i, j ) = 0, AF 2 = 1.5 β

(18)

Afterward, use the alteration factors AF1 and AF2 to compute the quantized coefficient values as: 𝐶1=2𝑘β+AF1, and 𝐶2=2𝑘β+AF2, where 𝑘 = floor(ceil(Ci,j(0,0) / 2β). Then, compute C ' i, j (0, 0) for embedding the watermark in Ci,j(0,0) using the following expression:

(

)

(

)

C if abs C i, j (0, 0) − C 2 < abs C i, j (0, 0) − C 1 C i′, j (0, 0) = 2 C 1 else

(19)

The modified value ∆𝐶𝑖,𝑗of the DC coefficient is then computedas follows: ∆Ci,j(0,0) = 𝐶′𝑖, (0,0) −𝐶𝑖,𝑗(0,0)

(20)

Finally, in order to obtain the watermarked image, ∆Ci,j(0,0)/8 is added to all the pixels of the block.

4.3. Watermark Extraction The proposed scheme uses blind extraction. The following steps are followed to extract the watermark from thewatermarked image. 1. Divide the watermarkedimage into non-overlapped 8 × 8 blocks.

9

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification

2. Obtain the DC coefficient 𝐶𝑖,𝑗(0,0) of each block of watermarked image directly using equation (6). 3. Compute the encrypted watermark that was embedded in the watermarked imageusing the quantification step β as follows: Ci, j (0, 0) , 2 e𝑤(𝑖,𝑗) = 𝑚𝑜𝑑 ceil β

(21)

4. Use same key K, at the receive for successful decryption of watermark

4.4. Proposed System Evaluation The proposed system is analysed for these parameters using various subjective and objective image quality metrics. Some of the objective image quality metrics used to evaluate the scheme have been defined below. It is pertinent to mention that Cj,k and Wj,k respectively represent cover image and its watermarked version. P and Q are number of pixels in rows and columns while as wm and wme respectively denote embedded and extracted watermarks. A. Peak Signal to Noise Ratio (PSNR) PSNR = 10log

(2n − 1)2 (255)2 = 10log MSE MSE

(22)

B. Mean Square Error (MSE) P

MSE =

Q

1 (C j,k − Wj,k )2 PQ ∑∑ j=1 k =1

(23)

C. Normalized cross co-relation (NCC)

NCC =

P

Q

j=1

k =1

∑ ∑

(C j,k * Wj,k )

P

Q

j=1

k =1

∑ ∑

C2 j,k

Bit Error Rate (BER) =

10

1 PQ

P Q ∑∑wm (i, j ) ⊕ wme (i, j ) ×100 i =1 j =1

(24)

(25)

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification

5. EXPERIMENTAL RESULTS AND DISCUSSIONS Digital image watermarking systems are supposed to satisfy some important requirements like imperceptibility, robustness and security.

5.1. Imperceptibility Analysis Imperceptibility in a wartermarking system refers to the fact that addition of watermark shoud to add a perceptible artifacts to the original image. This section presents various subjective and objective experimental results obtained from the imperceptivity point of view of the proposed system. It is to be noted that the proposed scheme is tested on a number of standard grey scale test image of size (512 x512), while as size of the embedded watermark is 64× 64. The investigation has been carried out on Intel core i5-560M, 2.66 GHz processor with 4GB RAM using MATLAB 7 running on Windows platform. Figure 3 shows various original images and their respective watermarked imagesobtained using proposed scheme, while as objective metrics obtained are presented in Table 1. It is evident from the subjective quality of the watermarked images that the proposed technique is capable of producing highly imperceptible watermarked images From Table 1,it is seen that average PSNR obtained for the test images is 42.82 dB and NCC is 1.0000 showing that high quality imperceptible images are produced by the proposed scheme. A comparison of the PSNR of the proposed scheme with various state- of-art schemes for test image Lena is shown in Figure 4.

Figure 3. Original images and their corresponding watermarked versions

11

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification

Table 1. Various Image quality metrics for watermarked images Image

PSNR (dB)

NCC

Lena

42.83

1.0000

Baboon

42.89

1.0000

Peppers

42.84

1.0000

Plane

42.72

1.0000

Average

42.82

1.0000

Figure 4. Comparison of proposed scheme with state-of art- for “Lena”

As shown the proposed scheme outperforms all the techniques under comparison. A further comparison of the proposed scheme with various a latest state-of-art schemesas depicted in Table 2, whichshows the superiority of the proposed scheme.

5.2 Robustness Analysis Robustness refers to ability of a watermarking scheme to withstand various attacks. The watermarked images are subjected to various image processing and geometric attacks to investigate the robustness of the proposed scheme. Figure 5 shows subjective quality of watermarked images after different attacks, for test image Lena while as Table 3 shows various objective quality parameters obtained when watermarked Lena is subjected to various attacks alike, noise addition, filtering, histogram equalization and JPEG compression.

12

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification

Table 2. Comparison of proposed scheme with (Parah et al, 2016) PSNR (dB)

Image

(Parah et al, 2016)

Proposed

Lena

41.27

42.83

Baboon

41.21

42.89

Peppers

41.84

42.84

Plane

41.17

42.72

Average

41.37

42.82

Figure 5. Subjective quality of ‘Lena’ for various attacks

13

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification

Table 3. Various Image quality parameters after different attacks on watermarked ‘Lena’ PSNR (dB)

BER

NCC

No Attack

42.8312

0

1.0000

Median Filtering (3x3)

34.7958

0.0430

Attack

Table 4. Various Image quality parameters after different attacks on watermarked ‘Baboon’ PSNR (dB)

BER

NCC

No Attack

42.8945

0

1.0000

0.9886

Median Filtering (3x3)

25.0021

0.0700

0.9501

25.5055

0.3098

0.9321

Attack

Salt and Pepper Noise (d=0.01)

25.1388

0.1531

0.9718

Salt and Pepper Noise (d=0.01)

Histogram Equalization

19.0975

0.4836

0.8859

Histogram Equalization

16.2279

0.5020

0.8632

Gaussian Noise (v=0.0002)

29.6893

0.0164

0.9963

Gaussian Noise (v=0.0002)

29.7877

0.1489

0.9736

JPEG (Quality Factor)

JPEG (Quality Factor)

10

29.5628

0.0188

0.9477

10

24.2698

0.5479

0.8034

20

29.5628

0.0773

0.9505

20

26.0754

0.3052

0.9014

30

30.5493

0.0674

0.9894

30

27.1856

0.1152

0.9727

40

31.1212

0.0071

0.9994

40

28.0265

0.0071

0.9924

50

31.5609

0.0012

0.9997

50

28.7920

0

1.0000

60

31.9095

0

1.0000

60

29.6154

0

1.0000

70

32.3640

0

1.0000

70

31.5067

0

1.0000

80

32.9650

0

1.0000

80

32.8423

0

1.0000

Low Pass Filtering (3x3)

30.7020

0.0583

0.9607

Low Pass Filtering (3x3)

24.4918

0.1387

0.9665

Sharpening (3x3)

25.8487

0.6943

0.7478

Sharpening (3x3)

21.6806

0.4309

0.8856

1

42.8943

0.1007

0.9889

10

42.6945

0.1118

0.9848

45

42.8945

0.1703

0.9801

Centre

42.8933

0.0928

0.9968

Top Left Corner

42.8945

0.2039

1.0000

Bottom left corner

42.8712

0.2332

1.0000

Top right Corner

42.8815

0.2126

0.9996

Bottom right corner

42.8777

0.2364

0.9960

Rotation (Degrees)

Rotation (Degrees)

1

42.4311

0.0590

0.9882

10

42.8314

0.0640

0.9886

45

42.0314

0.1685

0.9880

Centre

42.8314

0.0918

0.9991

Top Left Corner

42.6780

0.1678

0.9993

Bottom left corner

42.2345

0.2322

0.9996

Top right Corner

42.7717

0.0918

0.9989

Bottom right corner

42.8022

0.1254

0.9979

Cropping

Cropping

It is pertinent to mention that BER and NCC have been computed between original and extracted watermark, while as PSNR has been computed between Original and Watermarked images.The results in Figure 5 and Table 3 depict that the proposed scheme is robust to the image processing attacks. Tables 4, 5 and 6 respectively show the performance indices of the proposed scheme for test images ‘Baboon’, ‘Peppers’ and ‘Plane’; respectively.

14

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification

Table 5. Various Image quality parameters after different attacks on watermarked Pepper PSNR (dB)

BER

NCC

No Attack

42.833

0

1.0000

Median Filtering(3x3)

34.5423

0.0291

Attack

Table 6. Various Image quality parameters after different attacks on watermarked ‘Plane’ PSNR (dB)

BER

NCC

No Attack

42.7276

0

1.0000

0.9945

Median Filtering (3x3)

34.0620

0.0735

0.9803

25.0672

0.2769

0.9416

Attack

Salt and Pepper Noise (d=0.01)

25.1283

0.2874

0.9393

Salt and Pepper Noise (d=0.01)

Histogram Equalization

20.5714

0.4827

0.8558

Histogram Equalization

13.4360

0.5393

0.8509

Gaussian Noise (v=0.0002)

29.8138

0.1423

0.9754

Gaussian Noise (v=0.0002)

29.7704

0.1436

0.9708

JPEG (Quality Factor)

JPEG (Quality Factor)

15

31.2246

0.5374

0.8504

15

31.1593

0.4380

0.8020

25

32.6312

0.2197

0.9610

25

33.0126

0.1912

0.9669

35

33.3813

0.0598

0.9898

35

34.1132

0.0635

0.9945

45

33.8969

0.0095

1.0000

45

34.8263

0.0059

0.9994

55

34.3289

0

1.0000

55

35.4286

0

1.0000

65

34.8219

0

1.0000

65

36.1732

0

1.0000

75

35.7234

0

1.0000

75

37.0012

0

+1.0000

85

36.1135

0

1.0000

85

37.6519

0

1.0000

Low Pass Filtering (3x3)

31.2781

0.0754

0.9581

Low Pass Filtering (3x3)

30.7850

0.1052

0.9442

Sharpening (3x3)

26.8880

0.3079

0.9198

Sharpening (3x3)

24.8422

0.3030

0.9191

Rotation (Degrees)

Rotation (Degrees)

1

42.6716

0.0332

0.9919

1

42.8192

0.0542

0.9889

10

42.8335

0.0632

0.9889

10

42.7276

0.0713

0.9841

45

42.6331

0.1589

0.9938

45

41.1216

0.1848

0.9819

Cropping

Cropping

Centre

42.8335

0.0938

0.9950

Centre

42.7276

0.0993

0.9956

Top Left Corner

42.8515

0.02039

1.0000

Top Left Corner

42.1654

0.2039

1.0000

Bottom left corner

42.8003

0.02017

0.9996

Bottom left corner

42.3213

0.2334

1.0000

Top right Corner

42.6315

0.02014

0.9992

Top right Corner

42.1843

0.2114

0.9992

Bottom right corner

42.3641

0.0364

0.9660

Bottom right corner

42.0907

0.0371

0.9662

5.3. Brief Discussions on Robustness A brief discussion on robustness of this system to various attacks has been carried out for the system and is presented below. A. Median Filtering: We have subjected the watermarked images is subjected to Median Filtering attack with a kernel of (3×3).The BER to this attack varies from 2% to 7% for the various test images used for evaluating the scheme, showing that the proposed scheme shows a good degree of robustness to this attack.

15

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification

B. Salt and Pepper attack: The watermarked images obtained using proposed scheme have been attacked with salt and pepper noise of (density=0.01). The observed values of BER for various test images vary from 27% to 30%. C. Histogram Equalization: The performance of the proposed scheme is not as good for this attack compared to above mentioned attacks. This is substantiated by high values of BER obtained for various images. D. Additive White Gaussian Noise (AWGN): The proposed scheme has been tested for Gaussian noise attack (variance=0.0002). It has been observed that BER for this attack varies for 14% to 16% for various used images. Thus the scheme is robust to this attack. E. Low Pass Filtering: We have subjected our scheme to Low Pass Filtering attack with filter kernel size of (3×3). It has been seen that the scheme is robust to this attack as the observed BER varies from 5% to 13%. F. Sharpening: The scheme has been subjected to sharpening (3×3) attack. It has been observed that the scheme is less robust to this attack as the observed BER ranges from 30% to 43%. G. Rotation: It has been observed that the proposed scheme is also robust to rotation attack. The observed BER values show that, robustness is highest for lower rotational values while it decreases as rotation increases. As observed the BER varies from 3% to 18% as angle of rotation is increased from 1 degree to 45 degree. H. JPEG Compression: JPEG is one of the mostly used compression standards,used to compress multimedia content like images and videos. It is generally used to save memory space for storing a multimedia object or to save bandwidth requirements. As such it forms on of the most important attacks (though unintentional at times) that a multimedia object like a digital image undergoes. We have exclusively tested our scheme for this attack for varying Quality Factor of JPEG compression. The Quality factor has been varied from 10 to 80. The experimental results show that the proposed scheme is highly robust to this attack. From Table 3 we observe that BER varies from 1% to 7% for quality factor values in the range of 10 to 30. It is pertinent to mention that smaller the quality factor higher the compression. However for any value of Quality factor above 50 the BER reduce to zero. From Table 4 it is observed that BER is 0 for quality factor of 50 or above. Similarly Tables 5 and 6 show that BER is zero for quality factor of 55 and above. It is as such concluded that the proposed scheme is absolutely robust to JPEG compression for a quality factor of 50 and above. This is one of the prominent advantages of the proposed scheme. Various robustness parameters of the proposed technique are comparedwith that of (Parah et al, 2016) and (Das et al., 2014) . In addition, the results have been depicted in Tables 7 and 8 as shown below. From the tabulated results it is quite evident that proposed scheme is more robust compare various schemes under comparison. Generally, security using several techniques such as watermarking is an interesting process that applied in several applications as reported in (Dey et al, 2013a; Parah et al, 2015a; Bose et al, 2014; Dey et al, 2013b; Dey et al, 2012a; Parah et al 2015a; Dey et al, 2012b; Parah et al, 2015b; Shabir et al, 2014b, Shabir et al, 2013b; Shabir et al, 2012b; Bhat et al, 2010; Bhat et al, 2009). The proposed approach establishes its efficiency, thus it can be applied with various applications.

16

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification

Table 7. Robustness comparison of proposed scheme with (Parah et al, 2016) for ‘Lena’ Attack

Normalized Cross-Correlation (NC) (Parah et al, 2016)

PSNR (dB)

Proposed

(Parah et al, 2016)

Proposed

Median Filtering (3×3)

0.9445

0.9886

33.52

34.79

Salt and Pepper noise (0.01)

0.8598

0.9798

24.9

25.13

Histogram Equalization

0.9665

0.8893

11

19.09

AWGN (0.0001)

0.9375

0.9963

29.69

30.52

Sharpening

0.99731

0.9778

25.69

25.84

Table 8. Robustness comparison of proposed scheme with (Parah et al, 2016) and (Das et al., 2014) for cropping attack for test Image ‘Lena’ CroppedArea

Normalized Cross-Correlation (NC) (Das et al, 2014)

(Parah et al, 2016)

Proposed

25% Top-left corner

0.9954

0.9986

0.9993

25% Top-right corner

0.9973

0.9980

0.9996

25%Bottom-left corner

0.9924

0.9989

0.9989

25%Bottom-right corner

0.9981

0.9980

0.9979

6. CONCLUSION This paper presented a robust and secure image watermarking scheme for grey scale images. The cover image was divided into 8×8 sub-blocks and instead of computing DC coefficient using DCT. TheDC coefficient of each block in spatial domain was computedand modified in to embed watermark. The quantum of change to be brought in various pixels of a block for embedding watermark bit depends upon DC coefficient of respective blocks, nature of watermark bit (0 or 1) to be embedded and the adjustment factor. Chaotic encryption was used for ensuring security of embedded watermark. Extensive experimental investigations were carried out for imperceptibility and robustness analysis of the scheme. Experimental investigations demonstrated that besides being highly secure,the proposed technique is robust to various signal processing and geometric attacks. The most important advantage of the system is that it is absolutely robust to JPEG compression for quality factor of 50 and above. The proposed scheme was compared with some state-of- art schemes in the field and comparison results show that our scheme performs better. In addition to above facts it is imperative to understand that the proposed scheme is computationally efficient as DC coefficient which holds the watermark information has been computed in pixel domain instead of using DCT on an image block. The robustness of the proposed scheme to histogram equalization and sharpening, however, is low. The future work aims at modifying the algorithm so that it can withstand these attacks as well.

17

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification

ACKNOWLEDGMENT The authors acknowledge the support rendered by University Grants Commission (UGC) Government of India under its Minor Research Project grant number 41-1340/2012(SR) for conduct of this work.

REFERENCES Bhat, G., Mustafa, M., Shabir, A., & Javaid, A. (2010). FPGA Implementation of Novel Complex PN Code Generator based data Scrambler and Descrambler. International journal of Science and Technology, 4(1), 125-135. Bhat, G., Shabir, A., & Javaid, A. (2009). VHDL Modelling and Simulation of Data Scrambler and Descrambler for secure data communication. Indian Journal of Science and Technology, 2(10), 41–43. Bhatnagar, G., & Raman, B. (2011). A new robust reference logo watermarking scheme. Multimedia Tools and Applications, 52(2), 621–640. doi:10.100711042-009-0433-2 Bhatnagar, G., & Wu, Q. (2015). A new robust and efficient multiple watermarking scheme. Multimedia Tools and Applications, 74(19), 8421–8444. doi:10.100711042-013-1681-8 Bose, S., Chowdhury, S. R., Sen, C., Chakraborty, S., Redha, T., & Dey, N. (2014). Multi-thread video watermarking: A biomedical application. Proceedings of International Conference on Circuits, Communication, Control and Computing (I4C) (pp. 242-246). Chen, R., Luo, Y., Lan, Y., & Alsharif, R. (2013). A new robust digital image watermarking algorithm based on singular value decomposition and independent component analysis. J. Com. Inf. Tech., 8(5), 530–537. Cintra, J., Dimitrov, S., Oliveira, M., & Campello, M. (2009). Fragile watermarking using finite field trigonometrical transforms. Signal Processing Image Communication, 24(7), 587–597. doi:10.1016/j. image.2009.04.003 Cox, I., Miller, L., Bloom, A., Fridrich, J., & Kalker, T. (2008). Digital watermarking and steganography. Elsevier. Das, C., Panigrahi, S., Sharma, V., & Mahapatra, K. (2014). A novel blind robust image watermarking in DCT domain using inter-block coefficient correlation. Int. J. Electron. Commun, 68(3), 244–253. doi:10.1016/j.aeue.2013.08.018 Dey, N., Das, P., Chaudhuri, S. S., & Das, A. (2012c). Feature analysis for the blind-watermarked electroencephalogram signal in wireless telemonitoring using Alattar’s method. Proceedings of the Fifth International Conference on Security of Information and Networks (pp. 87-94). 10.1145/2388576.2388588 Dey, N., Das, P., Roy, A. B., Das, A., & Chaudhuri, S. S. (2012a). DWT-DCT-SVD based intravascular ultrasound video watermarking. Proceedings of Information and Communication Technologies (pp. 224–229). WICT.

18

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification

Dey, N., Mukhopadhyay, S., Das, A., & Chaudhuri, S. S. (2012b). Analysis of P-QRS-T components modified by blind watermarking technique within the electrocardiogram signal for authentication in wireless telecardiology using DWT. International Journal of Image. Graphics and Signal Processing, 4(7), 33–36. doi:10.5815/ijigsp.2012.07.04 Dey, N., Nandi, B., Das, P., Das, A., & Chaudhuri, S. (2013a). Retention of electrocardiogram features insignificantly devalorized as an effect of watermarking. Advances in Biometrics for Secure Human Authentication and Recognition, 23, 175–184. doi:10.1201/b16247-10 Dey, N., Samanta, S., Yang, X. S., Das, A., & Chaudhuri, S. S. (2013b). Optimisation of scaling factors in electrocardiogram signal watermarking using cuckoo search. International Journal of Bio-inspired Computation, 5(5), 315–326. doi:10.1504/IJBIC.2013.057193 Djurovic, I., Stankovic, S., & Pitas, I. (2001). Digital watermarking in the fractional Fourier transformation domain. Journal of Network and Computer Applications, 24(2), 167–173. doi:10.1006/jnca.2000.0128 Fazlali, H., Samavi, S., Karimi, N., & Shirani, S. (2016). Adaptive blind image watermarking using edge pixel concentration. Multimed Tools Appl. Doi:10.100711042-015-3200-6 Ghouti, L., Bouridane, A., Ibrahim, M., & Boussakta, S. (2006). Digital image watermarking using balanced multiwavelets. IEEE Transactions on Signal Processing, 54(4), 1519–1536. doi:10.1109/ TSP.2006.870624 Ghouti, L., Bouridane, A., Ibrahim, M., & Boussakta, S. (2006). Digital image watermarking using balanced multiwavelets. IEEE Transactions on Signal Processing, 54(4), 1519–1536. doi:10.1109/ TSP.2006.870624 Guo, J., Zheng, P., & Huang, J. (2015). Secure watermarking scheme against watermark attacks in the encrypted domain. Journal of Visual Communication and Image Representation, 30, 125–135. doi:10.1016/j.jvcir.2015.03.009 Kalra, G., Talwar, R., & Sadawarti, H. (2015). Adaptive digital image watermarking for color images in frequency domain. Multimedia Tools and Applications, 74(17), 6849–6869. doi:10.100711042-014-1932-3 Lai, C., & Tsai, C. (2010). Digital image watermarking using discrete wavelet transform and singular value decomposition. IEEE Transactions on Instrumentation and Measurement, 59(11), 3060–3063. doi:10.1109/TIM.2010.2066770 Lang, J., & Zhang, Z. (2014). Blind digital watermarking method in the fractional Fourier transform domain. Optics and Lasers in Engineering, 53, 112–121. doi:10.1016/j.optlaseng.2013.08.021 Lin, S., Shie, S., & Guo, J. (2010). Improving the robustness of DCT-based image watermarking gainst JPEG compression. Computer Standards & Interfaces, 32(1-2), 54–60. doi:10.1016/j.csi.2009.06.004 Liu, R., & Tan, T. (2002). An SVD-based watermarking scheme for protecting rightful ownership. IEEE Transactions on Multimedia, 4(1), 121–128. doi:10.1109/6046.985560 Liu, Y., & Zhao, J. (2010). A new video watermarking algorithm based on 1D DFT and Radon transform. Signal Processing, 90(2), 626–639. doi:10.1016/j.sigpro.2009.08.001

19

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification

Lu, W., Sun, W., & Lu, H. (2012). Novel robust image watermarking based on subsampling and DWT. Multimedia Tools and Applications, 60(1), 31–46. doi:10.100711042-011-0794-1 Ma, F., Zhang, J., & Zhang, W. (2012). A blind watermarking technology based on DCT do-main, Proc. of the IEEE Int. Conf. on Computer Science and Service System (pp. 398–401) Niansheng, L., Huajian, L., Huaiyu, D., Donghui, G., & Deming, C. (2015). Robust blind image watermarking based on chaotic mixtures. Nonlinear Dynamics, 80(3), 1329–1355. doi:10.100711071-015-1946-z Parah, S., Javaid, A., Farhana, A., & Bhat, G. (2015d). On the Realization of Robust Watermarking System for Medical Images. Proc. of 12th IEEE India International Conference (INDICON) on Electronics, Energy, Environment, Communication, Computers, Control (E3-C3), New Delhi, India (pp. 172-178). Parah, S., Javaid, A., Jahangir, A., Nazir, L., Farhana, A., & Bhat, G. (2015c). A high Capacity Data Hiding Scheme Based on Edge Detection and Even-Odd Plane Separation. Proceedings of 12th IEEE India International Conference (INDICON) on Electronics, Energy, Environment, Communication, Computers, Control (E3-C3), New Delhi, India (pp. 1-6) 10.1109/INDICON.2015.7443595 Parah, S., Javaid, A., Nazir, L., Farhana, A., & Bhat, G. (2015b). Information Hiding in Medical Images: A Robust Medical Image Watermarking System for E-Healthcare. Multimedia Tools and Applications. doi:10.100711042-015-3127-y Parah, S., Shazia, A., & Ayash, A. (2015a). Robustness analysis of a digital image watermarking technique for various frequency bands in DCT domain. Proceedings of international 1st IEEE international symposium on nanoelectronic and information systems iNiS, Indore India (pp. 71-76). 10.1109/iNIS.2015.41 Parah, S., Sheikh, J., Hafiz, M., & Bhat, G. (2014a). A secure and robust information hiding technique for covert communication. International Journal of Electronics, 102(8), 1253–1266. doi:10.1080/002 07217.2014.954635 Parah, S., Sheikh, J., Loan, N., & Bhat, G. (2016). Robust and blind watermarking technique in DCT domain using inter-block coefficient differencing. Digital Signal Processing, 53, 11–24. doi:10.1016/j. dsp.2016.02.005 Qingtang, S., Yugang, N., Qingjun, W., & Guorui, S. (2013). A blind color image watermarking based on DC component in the spatial domain. Optik (Stuttgart), 124, 255–6260. Shabir, A., Javaid, A., & Bhat, G. (2012a). Data hiding in ISB planes: A high capacity blind stenographic technique. Proc. of IEEE sponsored Intl. Conference INCOSET ‘12, Tamilnadu, India (pp. 192-197). Shabir, A., Javaid, A., & Bhat, G. (2012b). On the realization of a secure, high capacity data embedding technique using joint top-down and down- top embedding approach. Elixir Comp. Sci. & Engg., 49, 10141–10146. Shabir, A., Javaid, A., & Bhat, G. (2013a). On the Realization of a spatial Domain Data Hiding Technique based on Intermediate Significant Bit Plane Embedding (ISBPE) and Post Embedding Pixel Adjustment (PEPA). Proceedings of IEEE International Conference on Multimedia Signal Processing and Communication Technologies IMPACT ‘13, AMU, Aligargh, India (pp. 51-55)

20

Realization of a New Robust and Secure Watermarking Technique Using DC Coefficient Modification

Shabir, A., Javaid, A., & Bhat, G. (2013b). Data hiding in color images: A high capacity data hiding technique for covert communication. Computer Engineering and Intelligent Systems, 4, 113–118. Shabir, A., Javaid, A., & Bhat, G. (2013c). High capacity data embedding using joint intermediate significant bit and least significant technique. International Journal of Information Engineering and Applications, 2, 1–11. Shabir, A., Javaid, A., & Bhat, G. (2014a). A secure and efficient spatial domain data hiding technique based on pixel adjustment. American Journal of Engineering and Technology Research, 14(2), 38–44. Shabir, A., Javaid, A., & Bhat, G. (2014b). Data hiding in scrambled images: A new double layer security data hiding technique. Computers & Electrical Engineering, 40(1), 70–82. doi:10.1016/j.compeleceng.2013.11.006 Shabir, A., Javaid, A., & Bhat, G. (2014c). Fragility Evaluation Of Intermediate Significant Bit Embedding (ISBE) Based Digital Image Watermarking Scheme For Content Authentication. Proc. of International Conference on Electronics, Computers and Communication, Banglore, India (pp. 198-203) Shabir, A., Javaid, A., & Bhat, G. (2015). Hiding in encrypted images: A three tier security data hiding system. Multidimensional Systems and Signal Processing. doi:10.100711045-015-0358-z Shabir, A., Javaid, A., & Bhat G. (2012c). On The Realization Of Secure And Efficient Data Hiding System using ISB & LSB Technique. Engineering e-Transaction, 7(2), 48-53. Shih, F., & Wu, S. (2003). Combinational image watermarking in the spatial and frequency domains. Pattern Recognition, 36(4), 969–975. doi:10.1016/S0031-3203(02)00122-X Shih, Y. (2008). Digital watermarking and steganography: fundamentals and techniques. FL: CRC Press. doi:10.1007/978-3-540-92238-4 Tsai, M. (2011). Wavelet tree based digital image watermarking by adopting the chaotic system for security enhancement. Multimedia Tools and Applications, 52(2–3), 347–367. doi:10.100711042-010-0475-5 Wang, S., Zheng, D., Zhao, J., Tam, J., & Speranza, F. (2007). An image quality evaluation method based on digital watermarking. IEEE Transactions on Circuits and Systems for Video Technology, 17(1), 98–105. doi:10.1109/TCSVT.2006.887086

This research was previously published in the Journal of Global Information Management (JGIM), 25(4); edited by Zuopeng (Justin) Zhang; pages 80-102, copyright year 2017 by IGI Publishing (an imprint of IGI Global).

21

22

Chapter 2

Chaotic Function Based ECG Encryption System Butta Singh Guru Nanak Dev University, India Manjit Singh Guru Nanak Dev University, India Dixit Sharma Guru Nanak Dev University, India

ABSTRACT Remote health-care monitoring systems communicate biomedical information (e.g. Electrocardiogram (ECG)) over insecure networks. Protection of the integrity, authentication and confidentiality of the medical data is a challenging issue. This chapter proposed an encryption process having a 4-round five steps -encryption structure includes: the random pixel insertion, row separation, substitution of each separated row, row combination and rotation. Accuracy and security analysis of proposed method for 2D ECG encryption is evaluated on MIT-BIH arrhythmia database.

INTRODUCTION The technology advancements in health care systems have dramatically increased the number of elderly patients. Remote health care monitoring of patients can decrease the traffic at specialized medical centers and provide reliable emergency services. The applications of remote healthcare technologies have also reduced the medical costs as well. In remote health care monitoring, body sensors acquire biological signals and other physiological parameters of the patient. The recorded signals and confidential side information or any urgent alerts are sent to the specialized hospital servers or medical cloud via the Internet. The security and privacy threats as well as crucial biomedical data integration issues are introduced with internet as a communication channel. Secure transmission of confidential biomedical data has become a common interest in both research and applications (Leeet al., 2008; Li et al., 2013; Hu et DOI: 10.4018/978-1-7998-1763-5.ch002

Copyright © 2020, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Chaotic Function Based ECG Encryption System

al., 2007). Accordingly, it is essential to employ a security protocol which will have powerful information security. One method to protect information from unauthorized eavesdropping is to use an encryption technique. The encryption is the process by which the information is transformed into intelligible form to construct the encrypted data/cipher data. Decryption is the process to reconstruct the original information from encrypted data. An electrocardiogram (ECG) is an important physiological signal required to transmit in remote health care system used not only to analyze cardiac diseases, but also to provide crucial biometric information for identification and authentication. The ECG signal which monitors the electrical activity of heart is usually characterized by its various set points (P, QRS, T) and intervals (PR interval, QT interval and RR interval) that reflects the rhythmic electrical depolarisation and repolarisation of atria and ventricles (Singh et al., 2014). With an ECG signal, various arrhythmias, degree of myocardial damage and the structure of the atrium and ventricle can also be analyze and identified. While transmitting biomedical information such as ECG through the internet, protection of patient’s privacy and confidentiality is a challenging issue (Jero et al., 2015). The methods of computer software should guarantee the information security on the server and inside the communication channels. Several researchers have proposed various security protocols to secure patient confidential information (Enginet al., 2005; Ibaida et al., 2013). The Encryption algorithms based techniques are commonly used to secure data during the communication and storage. As a result, the final data will be stored in encrypted format (Wang et al., 2010; Maglogiannis et al., 2009). In 1998, Fridrich proposed the chaos-based approach for image encryption (Fridrich 1998), since then there have been increasing researches on chaotic encryption techniques. Chaos based algorithms are developed and considered as the core of encryption processes due to ergodicity, mixing property, the high sensitivity of chaotic systems to parameters and initial conditions (Zhu et al., 2011; Fu et al., 2011; Zhu et al., 2012). Recently, conventional logistic map and tent map based1D chaotic maps, and coupled map lattice based 2D chaotic maps have been developed for substitution-only encryption methods (Soma et al., 2013; Radwan et al., 2016). Chaos-based algorithms have shown exceptionally superior properties in aspects such as security, speed and complexity and computational cost. Many researchers have proposed ECG signal processing techniques by treating 1D ECG signal as a 2D image and exploiting the inter- and intra-beat correlations by encoder (Chou et al., 2006; Wang et al., 2008). The “cut and align beats approach and 2D DCT” and “period normalization and truncated SVD algorithm” are available preprocessing techniques to get good compression results in ECG (Wei et al., 2001; Lee et al., 1999). This kind of preprocessing are also often associated with the use of state-of-the-art image encoders, like JPEG2000. In (Chou et al., 2006), the authors proposed a lousy compression technique based on converting the 1D ECG signal into 2D ECG image. A period sorting preprocessing technique was introduced, which consists of a length-based ordering of all periods. The authors exploited inter and intra-beat dependencies to compress irregular ECG signals. The technique is based on the supposition that periods with similar lengths tend to be highly correlated, which is not a very strong assumption and may not be valid for pathological ECG signals. Another preprocessing technique consists of QRS detector, period length normalization, period preprocessing and image transform was proposed in (Filho et al., 2008). This chapter introduces a simple and efficient chaotic system approach using a combination of two existing 1D chaotic maps to encrypt 2D ECG signal. Security analysis reveals the performance of proposed method for 2D ECG encryption. ECG is encrypted in a lossless manner so that after reconstruction there will be zero difference between the original and the reconstructed ECG signal. 23

Chaotic Function Based ECG Encryption System

MATERIAL AND METHODS In the emerging field of medical engineering, research subjects like cardiac arrhythmia detection, heart rate variability, cardiovascular and pulmonary dynamics, artificial intelligence based medical decision support and ECG compression, etc. are of major interest (Singh et al., 2012a; Singh et al., 2012b). The Massachusetts Institute of Technology (MIT) supplies some valuable resources for such research projects. These resources include databases containing recorded physiological signals and software for analyzing, viewing and creating such recordings. In the present study, ECG samples from Arrhythmia Database (http://ecg.mit.edu) have been taken for investigation of the efficiency of the proposed method. The Database contains 48 half-hour excerpts of two-channel ambulatory ECG recordings, obtained from 47 subjects studied by the BIH Arrhythmia Laboratory. The 11-bit resolutions over 10 mV range with sampling frequency of 360 Hz per sample channel were recorded.

PROPOSED TECHNIQUE 1D ECG to 2D ECG Signal Formation of 2D ECG signal broadly consist steps: QRS detection, ECG segmentation, pre-processing, and transformation (Chou et al., 2006; Filho et al., 2008). First stage of process is the acquisition of ECG data from standard MIT-BIH arrhythmia database, for comparative analysis and validation of proposed technique. Details of this database are already discussed in section 2. The second stage involved QRS detection and R peak based ECG segmentation. The peaks of QRS complex were detected to identify each RR interval and to map 1D ECG signal to 2D ECG image, Many QRS detection algorithms have been proposed in literature and in the present work the RR interval time series were estimated by the Tompkins method proposed in (Pan et al., 1985)for its simple implementation and high detection accuracy In ECG segmentation, ECG samples from one R peak to next R peak are retained in one segmented block and each block (Row wise) is vertically stacked to form 2D ECG image(Chou et al., 2006). The segmentation and reassembling of ECG signal as an image was accomplished by choosing R peak as its delineation boundary, leaving half peak at each end of the row. Further the row oriented assembly was performed by retaining the ECG samples from one R-peak to the next R-peak. Then this 2D ECG will be encrypted by 2D chaotic method. Figure 1 shows the ECG samples for the record of 100, 117 and 119 from MIT-BIH arrhythmia ECG database. Figure 2 shows the result of 2D array of row wise stacking with their histograms in Figure 3. The resulted image was encoded through standard bitmap encoder which provides progressive quality. The resulted encrypted 2D ECG can be transmitted over cloud or a base station from remote place and can be reconstructed at receiving end. The first stage of reconstruction process is to split the side information and encrypted ECG from the merged data coming from communication channel. The next stage of reconstruction process is R peak based incorporation and estimation of raw ECG data.

Chaos Based ECG Encryption To investigate the proposed chaotic system in information security for ECG, using the Logistic Tent System as its example, we used 1D chaotic image encryption algorithm for ECG encryption. 24

Chaotic Function Based ECG Encryption System

Figure 1. First 2500 ECG samples of data 100, 117, and 119 of MIT-BIH arrhythmia database

The proposed algorithm has a 4-round-encryption structure. Each encryption round includes five steps: the random pixel insertion, row separation, substitution of each separated row, row combination and rotation with 90o in anticlockwise direction (Figure 4). The algorithm first inserts a random pixel in the beginning of each row in the original 2D ECG, separates each row into a 1D data matrix, applies a substitution process to change data values in each 1D matrix, combines all 1D matrices back into a 2D data matrix according to their row positions in the original 2D ECG, and then rotates the 2D matrix 90o anti clockwise. Repeating these processes four times obtains the final encrypted ECG (Figure 5). The proposed algorithm is able to transform original 2D ECG randomly into different noise-like encrypted images with excellent confusion and diffusion properties.

25

Chaotic Function Based ECG Encryption System

Figure 2. 2D ECG formation of complete data records 100, 117, and 119 of MIT-BIH arrhythmia database

Random ECG Sample Insertion One ECG sample with a random value is inserted in the beginning of each row in the original 2D ECG signal. Rand (x ) y =1 PE (x , y ) = E (x , y − 1) otherwise where E is the original 2D ECG with size of M X N; PE(x,y)is the processed ECG with size ofMX(N+1), 1≤x≤M, 1≤y≤N+1;Random function Rand(x) produces random numbers.

Chaotic Substitution The each row of PE is separated row by row into 1D array rowi(j)=PE(x,y) where rowi is the ith 1D row array with length of (N+1) The chaotic substitution process replaces the samples in each 1D row array rowi by the following equation:

26

Chaotic Function Based ECG Encryption System

Figure 3. Histogram of 100, 117, and 119 data records of MIT-BIH arrhythmia database

rowi ( j ) φi ( j ) = φ ( j − 1) ⊕ rowi ( j ) ⊕ Sk (i, j )X 1010 mod 256 i

(

)

j =1 otherwise

where ⊕ denotes the bit-level XOR operation, . is the floor function, and Sk(i,j) is the random sequence for the kth (k = 1, 2, 3, 4) encryption round, which is generated by the Logistic-Tent system(LTS) as defined by the following equation:

27

Chaotic Function Based ECG Encryption System

Figure 4. Description of proposed ECG encryption and decryption technique

i = 0, j = 0, k = 1 S1 (0, 0) i = 0, j = 0, k = 3 S 2 (M , 0) i = 0, j = 0, k = 2, 4 Sk −1(N , 0) Sk (i, j ) = i > 1, j = 0 LTS (r0 , Sk (i − 1, 0) i > 1, j > 0 LTS (rk , Sk (i, j − 1) where rk and Sk(0,0)are the parameter and initial value in the kth encryption round, respectively; S1(0,0), ro and rk are defined by users. We utilized the Logistic and Tent maps as seed maps to form the LTS system, as defined in the following equation: (rX (1 − X ) + (4 − r )X / 2) mod 1 X < 0.5 i n n n X n +1 = (rX n (1 − X n ) + (4 − r )(1 − X n ) / 2) mod 1 X i ≥ 0.5 where parameter ris with range of (0,4).

28

Chaotic Function Based ECG Encryption System

Row Combination After changing data values in each row matrix in the 1D substitution process, the row combination is an inverse process of the row separation and random insertion. It combines all 1D matrices back into a 2D matrix, and removes the first pixel in each row. The process is defined in the following equation: C(i,j) = ϕi(j+1) where C is the 2D matrix with size of M X N and j≤N.

Rotation The process is to rotate the 2D ECG matrix 90 degrees counter clockwise as following equation: E(i,j) = C(j, N – i +1) After the first encryption round, E is the feedback to the input of the random pixel insertion process. Final encrypted ECG obtained by four encryption rounds. In this algorithm, security keys are composed of six portions: the LTS parameter (r0) and initial value S1(0.0), the LTS parameters in each encryption round (r1,r2,r3,r4). In decryption stage, the authorized users should have correct security keys and follow the inverse procedures of encryption. The inverse1D substitution is defined in the following equation:

(

)

Ri ( j ) = Bi ( j − 1) ⊕ Bi ( j ) ⊕ Sk (i, j )X 1010 mod 256

Figure 5. Encrypted ECG of complete data records 100, 117, and 119 of MIT-BIH arrhythmia database

29

Chaotic Function Based ECG Encryption System

Figure 6. Histograms of encrypted 100, 117, and 119 data records of MIT-BIH arrhythmia database

RESULT AND DISCUSSION Having described the proposed encryption technique, we present in this section the performance evaluation and simulation parameters. The proposed method encrypt ECG signal with excellent confusion and diffusion properties and high security level. These can evade the confidential medical information from leakage.

30

Chaotic Function Based ECG Encryption System

Figure 7. Decrypted ECG of complete data records 100, 117, and 119 of MIT-BIH arrhythmia database with key K1(s1=0.6; r0=3.997; r1=3.99; r2=3.96; r3=3.77; r4=3.99)

Accuracy Analysis Percentage Root Mean Square Difference (PRD) It is a measure of acceptable fidelity and degree of distortion introduced during encryption and decryption algorithm N

PRD(%) = 100 ×

∑ (X (n ) − X (n )) n =1

2

s

r

N

∑ (X (n )) n =1

2

s

where Xs(n) and Xr(n) are the original and decrypted signal respectively of data length N. Proposed method encrypt the ECG signal lossless way (PRD=0) for all the data records of MIT-BIH database (Table 1). From juridical and clinical point of view, the use of lossless encryption/decryption process is very much important.

Histogram Analysis Histogram analysis shows the distribution of pixel values across the whole image where peaks for some specific gray scale value appear. For effectively encrypted images this distribution should be flat (Fuet al., 2011; Radwanet al., 2016). A visual investigation of the proposed method can be observed from the histograms of 2D ECG before and after the encryption (Figure 3 and Figure 6). After the encryption

31

Chaotic Function Based ECG Encryption System

Figure 8. Histograms of decrypted 100, 117, and 119 data records of MIT-BIH arrhythmia database with key K1(s1=0.6; r0=3.997; r1=3.99; r2=3.96; r3=3.77; r4=3.99)

process all histograms became similar to a uniform distribution, independent of the nature of the original histogram, as expected for a strong encryption scheme.Figure 6 shows all encrypted ECG signals are noise- like ones.

Security Analysis There are several statistical methods for evaluating the noise-like encrypted images, including the information entropy and correlation analysis. A good encryption should resist all kinds of known attacks, it should be sensitive to the secret keys, and the key space should be large enough to make brute-force attacks infeasible. Some security analysis has been performed on the proposed encryption scheme.

32

Chaotic Function Based ECG Encryption System

Table 1. Security analysis of proposed technique on MIT BIH arrhythmia database Correlation Coefficient (with -900 Rotation)

Correlation Coefficient (with +900 Rotation)

Original Information Entropy

Encrypted Information Entropy

Decrypted Information Entropy

PRD

100

0.006

-0.0014

3.9375

7.9914

3.9375

0

101

-0.0096

-0.0105

4.4612

7.993

4.4612

0

102

-0.0036

0.0071

4.6856

7.992

4.6856

0

103

-0.0022

-0.0144

4.6233

7.9922

4.6233

0

104

0.0005

-0.0034

4.8545

7.9928

4.8545

0

105

-0.0057

0.0082

4.2095

7.9947

4.2095

0

106

-0.0042

0.0068

4.7057

7.992

4.7057

0

107

0.0041

-0.0003

6.5215

7.9916

6.5215

0

Data

108

-0.002

0.0017

3.3504

7.995

3.3504

0

109

-0.0033

-0.0038

5.4018

7.9927

5.4018

0

111

0.0029

0.0054

4.9472

7.9916

4.9472

0

112

-0.002

0.0002

5.1214

7.9903

5.1214

0

113

0.0139

0.0085

5.2189

7.992

5.2189

0

114

0.0034

0.0022

4.3191

7.9922

4.3191

0

115

0.0071

-0.0037

4.9676

7.9928

4.9676

0

116

0.0079

0.0023

5.3086

7.9912

5.3086

0

117

0.0099

-0.0006

5.021

7.9913

5.021

0

118

0.0063

0.0035

5.5502

7.9933

5.5502

0

119

0.0037

0.0106

4.6586

7.9944

4.6586

0

121

-0.002

0.0033

5.3168

7.9919

5.3168

0

122

0.0038

0.0018

4.8485

7.9917

4.8485

0

123

0.0017

-0.0056

4.4171

7.9937

4.4171

0

124

-0.0034

-0.0082

4.8685

7.991

4.8685

0

200

-0.0065

0

3.7793

7.9951

3.7793

0

201

-0.0015

0.0043

3.551

7.9935

3.551

0

202

-0.0006

0.0135

4.1958

7.9935

4.1958

0

203

0.0031

-0.0044

2.5367

7.9973

2.5367

0

205

0.0128

-0.0003

3.7393

7.9916

3.7393

0

207

-0.004

-0.0082

3.7522

7.995

3.7522

0

208

0.0009

-0.0131

4.6145

7.9945

4.6145

0

209

0.0107

-0.0027

4.708

7.9913

4.708

0

210

-0.0043

0.0013

3.1032

7.9963

3.1032

0

212

0.011

0.0092

5.3862

7.993

5.3862

0

213

0.0013

-0.0015

5.8204

7.9927

5.8204

0

214

-0.0078

-0.0113

4.2264

7.9944

4.2264

0

215

-0.0061

-0.0028

3.8604

7.9948

3.8604

0

217

-0.0041

-0.0102

6.1748

7.9921

6.1748

0

continues on following page

33

Chaotic Function Based ECG Encryption System

Table 1. Continued Data

Correlation Coefficient (with -900 Rotation)

Correlation Coefficient (with +900 Rotation)

Original Information Entropy

Encrypted Information Entropy

Decrypted Information Entropy

PRD

219 220

0.0023

0.0033

0.0001

-0.0036

4.3868

7.994

4.3868

0

4.3864

7.9921

4.3864

0

221

-0.0024

-0.0029

3.632

7.9947

3.632

0

222

0.0028

0.0042

4.3512

7.9926

4.3512

0

223 228

0.0003

0.0033

3.4806

7.9954

3.4806

0

0.0024

-0.0073

3.6248

7.9957

3.6248

0

230

0.0027

-0.0031

4.9464

7.9912

4.9464

0

231

-0.0122

0.0034

3.7568

7.9942

3.7568

0

232

-0.0039

0.0021

2.618

7.9965

2.618

0

233

-0.0053

0.0016

4.9541

7.9945

4.9541

0

234

-0.0014

0.0032

4.8296

7.9911

4.8296

0

Information Entropy Information entropy (IE) is a measure of uncertainty in a random variable and evaluates the randomness of an image (Fuet al., 2011; Zhuet al., 2012). Higher the IE ofan image means the excellent random property. F −1

IE = ∑ P (L = l ) log 2 l =0

1 P (L = l )

where F is the gray level and P(L=l) is the percentage of pixels with value equal to l. For a gray scale image maximum IE is 8. Table 1 shows the IE of original, encrypted and decrypted 2D ECG signals of MIT BIH arrhythmia database. IE values of original and decrypted signal are similar for all the data records. Whereas IE values of encrypted signal are very high (>7.9) for all the records which shows the level of randomness in encrypted signals.

Correlation To verify the robustness of the proposed algorithm against statistic attacks, Pearson correlation coefficients of neighbouring (adjacent) pixels in the original and encrypted 2D ECG signals have been analyzed. The objective of encryption procedure is to transform highly correlated original images in to noise-like encrypted images with low correlations (Fuet al., 2011; Zhuet al., 2012). The correlation can be computed by corrxy = E[(x – μx)(y – μy)]/(σxσy) where μ and σ are the mean value and standard deviation, respectively.

34

Chaotic Function Based ECG Encryption System

Table 1 validate that the proposed method generated encrypted signal with sufficiently low correlation of adjacent pixels.

Security Key Analysis Six parameters of encryption algorithmr0, r1, r2, r3 and S1(0,0) form the key space. In the key sensitivity test, ECG data from samples 100, 117 and 119 (Figure 1) are encrypted using proposed algorithm with an initial key set (K1) as s1=0.6;r0=3.997;r1=3.99;r2=3.96;r3=3.77;r4=3.99; Figure 7 and 8 show the decrypted signals and corresponding histograms respectively for data samples 100, 117 and 119 with keys K1, and then another key set K2 is generated with a small change applied to r0while keeping all other parameters unchanged as s1=0.6;r0=3.9970000001;r1=3.99;r2=3.96;r3=3.77;r4=3.99; Figure 9 and 10 show the decrypted signals and corresponding histograms respectively for data samples 100, 117 and 119 with keys K2. It shows the high key sensitivity of the proposed method in both the encryption and decryption processes. Original signal can be reconstructed using correct key (K1) only. Even a small change in the security key (e.g. K2) will lead to the erroneous decryption process. Figure 9. Decrypted ECG of complete data records 100, 117, and 119 of MIT-BIH arrhythmia database with wrong key K2 (s1=0.6; r0=3.9970000001; r1=3.99; r2=3.96; r3=3.77; r4=3.99)

35

Chaotic Function Based ECG Encryption System

Figure 10. Histograms of decrypted 100, 117, and 119 data records of MIT-BIH arrhythmia database with wrong key K2 (s1=0.6; r0=3.9970000001;r1=3.99; r2=3.96;r3=3.77; r4=3.99)

CONCLUSION In view of E-health recognition and patient privacy, it is essential to encrypt biomedical signals. In particular, medical security is required to provide; confidentiality, integrity and authentication. In this chapter, we develop a chaos-based encryption system for applications to 2D ECG signals. Excellent diffusion and confusion properties of proposed method can resist the chosen-plaintext attack. Particularly, encrypted signals of the proposed algorithm are random, non- repeated and unpredictable, even using the same set of security keys. The algorithm can also withstand the data loss and noise attacks. While arithmetic operations based encryption approaches have to deal round-off errors, our algorithm recover exact original information after an encryption/decryption procedure. The novelty of this method is that the reconstructed signal is an exact replica of the original one and has an excellent key sensitivity.

36

Chaotic Function Based ECG Encryption System

REFERENCES Chou, H. H., Chen, Y. J., Shiau, Y. C., & Kuo, T. S. (2006). An effective and efficient compression algorithm for ECG signals with irregular periods. IEEE Transactions on Bio-Medical Engineering, 53(6), 1198–1205. doi:10.1109/TBME.2005.863961 PMID:16761849 Engin, M., Çıdam, O., & Engin, E. Z. (2005). Wavelet transformation based watermarking technique for human electrocardiogram (ECG). Journal of Medical Systems, 29(6), 589–594. doi:10.100710916005-6126-0 PMID:16235811 Filho, E. B. L., Rodrigues, N. M. M., da Silva, E. A. B., de Faria, S. M. M., da Silva, V. M. M., & de Carvalho, M. B. (2008). ECG signal compression based on Dc equalization and complexity sorting. IEEE Transactions on Bio-Medical Engineering, 55(7), 1923–1926. doi:10.1109/TBME.2008.919880 PMID:18595813 Fridrich, J. (1998). Symmetric ciphers based on two-dimensional chaotic maps. International Journal of Bifurcation and Chaos in Applied Sciences and Engineering, 8(6), 1259–1284. doi:10.1142/ S021812749800098X Fu, C., Lin, B., Miao, Y., Liu, X., & Chen, J. (2011). A novel chaos-based bit-level permutation scheme for digital image encryption. Optics Communications, 284(23), 5415–5423. doi:10.1016/j.optcom.2011.08.013 Hu, F., Jiang, M., Wagner, M., & Dong, D. (2007). Privacy-preserving telecardiology sensor networks: Toward a low-cost portable wireless hardware/software codesign. IEEE Transactions on Information Technology in Biomedicine, 11(6), 619–627. doi:10.1109/TITB.2007.894818 PMID:18046937 Ibaida, A., & Khalil, I. (2013). Wavelet-based ECG steganography for protecting patient confidential information in point-of-care systems. IEEE Transactions on Bio-Medical Engineering, 60(12), 3322–3330. doi:10.1109/TBME.2013.2264539 PMID:23708767 Jero, S. E., Ramu, P., & Ramakrishnan, S. (2015). ECG steganography using curvelet transform. Biomedical Signal Processing and Control, 22, 161–169. doi:10.1016/j.bspc.2015.07.004 Lee, H., & Buckley, K. M. (1999). ECG data compression using cut and align beats approach and 2-D transforms. IEEE Transactions on Bio-Medical Engineering, 46(5), 556–564. doi:10.1109/10.759056 PMID:10230134 Lee, W., & Lee, C. (2008). A cryptographic key management solution for hipaa privacy/security regulations. IEEE Transactions on Information Technology in Biomedicine, 12(1), 34–41. doi:10.1109/ TITB.2007.906101 PMID:18270035 Li, M., Yu, S., Zheng, Y., Ren, K., & Lou, W. (2013). Scalable and secure sharing of personal health records in cloud computing using attribute based encryption. IEEE Transactions on Parallel and Distributed Systems, 24(1), 131–143. doi:10.1109/TPDS.2012.97 Maglogiannis, I., Kazatzopoulos, L., Delakouridis, K., & Hadjiefthymiades, S. (2009). Enabling location privacy and medical data encryption in patient telemonitoring systems. IEEE Transactions on Information Technology in Biomedicine, 13(6), 946–954. doi:10.1109/TITB.2008.2011155 PMID:19171519

37

Chaotic Function Based ECG Encryption System

Pan, J., & Tompkins, W. J. (1985). A real-time QRS detection algorithm. IEEE Transactions on BioMedical Engineering, 32(3), 230–236. doi:10.1109/TBME.1985.325532 PMID:3997178 Radwan, A. G., Abd-El-Haleem, S. H. & Abd-El-Hafiz, S. K. (2016). Symmetric encryption algorithms using chaotic and non-chaotic generators: A review. Journal of Advanced Research, 7, 193–208. PMID:26966561 Singh, B., Sharma, D., Singh, M., & Singh, D. (2014). An improved ASCII character encoding method for lossless ECG compression. Advances in Biomedical Science and Engineering, 1(2), 1–11. Singh, B., & Singh, D. (2012a). Effect of threshold value r on multiscale entropy based heart rate variability. Cardiovascular Engineering and Technology, 3(2), 211–216. doi:10.100713239-012-0082-x Singh, B., Singh, D., Jaryal, A. K., & Deepak, K. K. (2012b). Ectopic beats in approximate entropy and sample entropy-based HRV assessment. International Journal of Systems Science, 43(5), 884–893. doi :10.1080/00207721.2010.543478 Soma, S., & Sen, S. (2013). A non-adaptive partial encryption of grayscale images based on chaos. Proc Technol, 10, 663–671. doi:10.1016/j.protcy.2013.12.408 Wang, H., Peng, D., Wang, W., Sharif, H., Chen, H., & Khoynezhad, A. (2010). Resource-aware secure ECG healthcare monitoring through body sensor networks. IEEE Wireless Communications, 17(1), 12–19. doi:10.1109/MWC.2010.5416345 Wang, X., Meng, J., Tai, S. C., Sun, C. C., & Yan, W. C. (2008). A 2-D ECG compression algorithm based on wavelet transform and vector quantization. Digital Signal Processing, 18(2), 179–188. doi:10.1016/j. dsp.2007.03.003 Wei, J., Member, S., Chang, C., Chou, N., & Jan, G. (2001). ECG data compression using truncated singular value decomposition. IEEE Transactions on Information Technology in Biomedicine, 5(4), 290–299. doi:10.1109/4233.966104 PMID:11759835 Zhu, C. (2012). A novel image encryption scheme based on improved hyperchaotic sequences. Optics Communications, 285(1), 29–37. doi:10.1016/j.optcom.2011.08.079 Zhu, Z., Zhang, W., Wong, K., & Yu, H. (2011). A chaos-based symmetric image encryption scheme using a bit-level permutation. Information Sciences, 181(6), 1171–1186. doi:10.1016/j.ins.2010.11.009

This research was previously published in the Handbook of Research on Healthcare Administration and Management edited by Nilmini Wickramasinghe; pages 205-221, copyright year 2017 by Medical Information Science Reference (an imprint of IGI Global).

38

39

Chapter 3

Biometric Image Security Using Chaos Algorithm Sugandha Agarwal Amity University, India O.P. Singh Amity University, India Deepak Nagaria Bundelkhand Institute of Engineering and Technology, India

ABSTRACT In this world of Advanced Technology, the Biometrics are proved to be a significant method for user identification. However, the use of biometric is not new, but these days, with the increase in multimedia applications, it has gained its popularity in analysing human characteristics for security purposes. Biometric Encryption using Chaos Algorithm is a technique used to make it more convenient to the user and to provide high level security. The most prominent physical biometric patterns investigated for security purposes are the fingerprint, hand, eye, face, and voice. In the proposed image encryption scheme, an external secret key of 160-bit is used. The initial conditions for the logistic map are derived using the external secret key. The results obtained through experimental analysis provide an efficient and secure way for real-time image encryption and transmission.

INTRODUCTION In recent years, the communication has become easier, but hackers are smarter than anyone can think of. It is more important to get aware of and use the best techniques for secure image transfer. With the escalation of information exchange across the Internet, and the storage of sensitive data on an open network, cryptography has become an increasingly important feature of computer security. In this paper, we have focused on biometric encryption using chaos algorithm. A biometric is defined as a unique, measurable, physical or biological trait for automatically recognizing or authenticating the identity of a human being. DOI: 10.4018/978-1-7998-1763-5.ch003

Copyright © 2020, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Biometric Image Security Using Chaos Algorithm

It includes data such as retina, iris, fingerprint, face, DNA, vein patterns, hand geometry, typing rhythm, mouse dynamics and voice. The main applications of biometrics are access controls, national ID card, passport control, border control, criminal investigation, and terrorist identification. Biometric is a very powerful tool for security purposes because of its property of uniqueness. Thus, biometric authentication can replace the use of passwords to secure a key. It provides a strong link between an individual and a claimed identity. This offers convenience and secure identity confirmation. If biometrics are used in most of the security systems for example bank locker systems, online transactions etc. then one can get a relief from remembering various passwords. It provides direct connection between the password and the user. A password is not tied to a user, the system running the cryptographic algorithm is unable to differentiate between the authorised user and an attacker who fraudulently acquires the password of an authorised user. As an alternative to password protection, biometric authentication provides a new mechanism for key security by using a biometric to secure the cryptographic key. Instead of entering a password to access the cryptographic key, the use of this key is guarded by biometric authentication. When a user wishes to access a secured key, he or she will be prompted to allow for the capture of a biometric sample. Biometric authentication is becoming the most popular and most reliable user authentication mechanism, even it is vulnerable to attacks.

LITERATURE REVIEW Several encryption algorithms (Maniccam & Bourbakis, 2001; Jiun-In & Cheng, n.d.; Gu & Han, 2006; Seyedzade et al., 2010; Zhou et al., 2014; Younes & Jantan, 2008; Sinha & Singh, 2003; Zeghid et al., 2007; Xiao & Zhang, 2006; Alsafasfeh & Arfoa, 2011) have been proposed including Lossless image compression and encryption using SCAN by S.S.Maniccam, N.G. Bourbakis (Maniccam & Bourbakis, 2001), Mirror-like image encryption algorithm by Jiun-In Guo, Jui-Cheng Yen (Jiun-In & Cheng, n.d), Image encryption based on hash function by Seyed Mohammad Seyedzade, Reza Ebrahimi Atani and Sattar Mirzakuchak (Seyedzade et al., 2010) and Image encryption using Block-Based Transformation Algorithm by Mohammad Ali, Bani Younes and Aman Jantan (Younes & Jantan, 2008), but Chaos-based encryption techniques (Gu & Han, 2006; Zhou et al., 2014; Alsafasfeh & Arfoa, 2011) are considered good for practical use as these techniques provide a good combination of speed, high security, complexity and computational power etc. Consequently, the traditional ciphers like AES, DES, and RSA etc. are not suitable for real time image encryption as these ciphers require a large computational time and high computing power. For real time image encryption, only those ciphers are preferable which take lesser amount of time and at the same time without compromising security. To ensure the privacy and security of the biometric system, many scholars have conducted researches to explore the possible risk of biometric network and feasible measurement to guarantee the security. Uludag et al. analysed the challenges involved in biometric application in authentication system and limitations of the biometric cryptosystems (Uludag et al., 2004). Many techniques have been introduced since to reduce the vulnerability of biometric data. Alok a Sinha and Kehar Singh introduced a technique for Image Encryption using Digital Signatures, Soutar et al. (1996), proposed an algorithm on biometric encryption. Alghamdi et al. (2010) states that image encryption cannot be used for large amount of data and high resolution images. Hao et al. (2005) presented a secure way to integrate iris biometric with cryptography. Chaos-based cryptography is the latest and efficient way to develop fast and secure cryptography for image encryption. The chaotic behaviour is the random behaviour of a nonlinear system 40

Biometric Image Security Using Chaos Algorithm

and the important characteristics of chaos is its extreme sensitivity to initial conditions of the system. In Baptista (1988), Baptista explained cryptography with chaos. A number of new techniques, extensions or improved versions of the earlier techniques, have been proposed in recent years. The main idea behind this paper is to develop a secure and reliable encryption technique using chaotic functions. In the proposed algorithm, we have used the Logistic-map for chaotic sequence generation. Then, a user intrinsic key is generated from the biometrics and used as the initial value of the chaotic sequence. The random sequence produced by the chaotic phenomenon is used to encrypt and secure the biometric image. In Section III, we discuss the step by step procedure of image encryption and in Section IV, the experimental analysis of the proposed image encryption scheme such as histogram, PSNR, MSE, Correlation Coefficient and key space analysis etc. to prove its security against the most common attacks. Finally, in Section V, we conclude the paper.

PROPOSED METHOD In this section, we discuss the step by step procedure used for encrypting as well as decrypting the biometric image. See the diagrams in Figure 1 and Figure 2.

Figure 1. Block diagram of encryption

41

Biometric Image Security Using Chaos Algorithm

Figure 2. Block diagram of decryption

Chaotic Function Chaotic functions are similar to the noise signal. Chaotic signal plays very important role in case of encryption because of their advantages as sensitivity to primary condition, high key space, control parameters, ergodicity, and randomness. The chaotic function logistic map equation (Zhou et al., 2014) is as follows: Xn+1= rXn (1−Xn)

(1)

where r can be a value between 0 and 4. The parameter r and initial value Xn may represent the key, and determines the future value Xn+1. When r is a value in between 3.6 to 4, it demonstrates a chaotic behaviour.

Key Generation A secret key of 160-bit is extracted from the user’s biometrics and used for generating the initial values for the formation of chaotic series. The key is generated based on the following steps: • • • • •

Obtain processed biometric image; Divide the image into four equal parts; Select the first five pixels from each part of the image to form initial value; These selected five pixels are then used as encryption key to encrypt the part of image; Initial value of logistic map function can be determined by using following equation:

P = [P1, P2, P3, P4, P5] (Decimal) •

42

Following equation is then used to convert P into ASCII number as follows:

Biometric Image Security Using Chaos Algorithm

B = [P1,1, P1,2, P1,3, . . ., P2,1, P2,2, . . ., P5,7, P5,8] (ASCII) (Pareek et al., 2006) •

Equation is used to determine initial value of chaotic map function as follows:

U ok =

P1,1 × 239 + P1,2 × 238 + … + P2,1 × 231 + … + P5,7 × 21 + P5,8 × 20 240

(2)

For each part of plain image step b is repeated. For encrypting pixels in each part of plain image following equation is used: New Value=round (Uik ×255) ⊗ Old Value

(3)

Encryption and Decryption Algorithm The steps of proposed encryption and decryption algorithms are presented below: 1. From each channel of image, extract the key as explained in the Key Generation algorithm; 2. The initial values for the chaotic map are obtained from the extracted key as explained above; 3. Encrypt each shuffled pixel by XORing with respective chaotic sequence generated. An encrypted image is obtained after performing the above steps. To decrypt the encrypted image, repeat the same steps as in encryption by reversing the process.

EXPERIMENT ANALYSIS • •

Histogram: The histogram of the original image (Figure 3) and encrypted image (Figure 4) is shown in this section; Correlation Coefficient: Correlation between two adjacent pixels is able to break the high correlation among the pixels. Therefore, the proposed technique is robust against statistical attacks. The lesser the correlation coefficient, the best is the ciphered image. We have calculated the correlation coefficient by using the following formula:

Correlation = ρ =

cov (X ,Y ) σx σy

(4)

where Covariance is obtained by this formula: Covxy =

Σ (x − x ) (y − y )

(n − 1)

=

Σxy − nxy

(n − 1)

(5)

43

Biometric Image Security Using Chaos Algorithm

• • •

Peak signal-to-noise ratio (PSNR) is the ratio between the maximum possible power of a signal and the power of corrupting noise. The higher PSNR means good quality of image. In this case, the PSNR value is 7.2393 and it is high enough which shows the image possess good quality; MSE: The Mean Square Error is the average squared difference between a reference image and a reconstructed (encrypted) image. The MSE is less i.e., 1.2279e+04, so we can conclude that the obtained ciphered image is the best image; Key Space: The key space size is the total number of different keys that can be associated with the encryption framework. For a secure encryption framework, the key space should be large enough to make the system vulnerable to brute-force attacks. From the cryptographic point of view, the size of the key space should not be smaller than 2100. In the proposed technique, we use 160-bit space key, therefore the key space is 2160. So, the-e key space is large enough to resist the bruteforce attacks.

Simulation The schematic flow chart in Figure 5 represents how the biometric image is acquired from the source, encrypted and decrypted using the chaotic function, for the use in various applications. See also Table 1.

Figure 3. Histogram of original image

Figure 4. Histogram of encrypted image

44

Biometric Image Security Using Chaos Algorithm

Figure 5. Simulation result

45

Biometric Image Security Using Chaos Algorithm

Table 1. Corresponding values obtained between original and encrypted image Parameters

Values

Correlation coefficient of original image

0.8971

Correlation coefficient of encrypted image

0.6490

Mean Square Error

1.6892e+004

Peak Signal-To-Noise Ratio

5.8540

Table 2. MSE and PSNR values MSE and PSNR Found Between

MSE Values

PSNR Values

Original and encrypted image

1.6892e+004

5.8540

Encrypted and decrypted image

1.6892e+004

5.8540

0

infinite

Original and decrypted image

From Table 2, MSE value 0 shows that reconstructed image is equal to the original image and there is no difference between them. Therefore, mean square error is zero. And PSNR being inversely proportional to MSE it is finite.

CONCLUSION In this paper, we have proposed a chaotic algorithm for enhancing the security of biometric data using the chaotic map. The importance of biometric approaches for authentication is increasing day by day because of the advantages such as security, accuracy, reliability, usability, and friendliness, etc. Some significant parameters in analysing the biometric data are lesser correlation coefficient, low MSE, high PSNR and large key space increase the robustness of the proposed cryptosystem. The experimental analysis including histogram, PSNR, MSE and Key Space shows that the proposed image encryption scheme provides an efficient and secure way for real-time image encryption as well as decryption process using chaos logistic map. The proposed system is not limited to colour biometric data and is extended to secure grey scale biometric data also.

REFERENCES Alghamdi, A. S., & Ullah, H. (2010). A Secure Iris Image Encryption Technique Using Bio-Chaotic Algorithm. International Journal of Computer and Network Security, 2(4), 78–84. Alsafasfeh, Q. H., & Arfoa, A. A. (2011). Image Encryption Based on the General Approach for Multiple Chaotic Systems. Journal of Signal and Information Processing, 2(3), 238-244. Baptista, M. S. (1988). Cryptography with chaos. College Park: Institute for Physical Science and Technology: University of Maryland. Gu, G., & Han, G. (2006). An Enhanced Chaos Based Image Encryption Algorithm. Proceedings of the IEEE First International Conference on Innovative Computing, Information and Control. Hao, F., Anderson, R., & Daugman, J. (2005). Combining cryptography with biometrics effectively (Tech. Rep. No. 640). Cambridge, United Kingdom: University of Cambridge, Computer Laboratory.

46

Biometric Image Security Using Chaos Algorithm

Jiun-In, G., & Cheng, Y.J., A new mirror-like image encryption algorithm and its VLSI architecture. Department of Electronics Engineering National Lien-Ho College of Technology and Commerce. Maniccam, S. S., & Bourbakis, N. G. (2001). Lossless image compression and encryption using SCAN. Pattern Recognition, 34(6), 1229–1245. doi:10.1016/S0031-3203(00)00062-5 Pareek, N.K., Patidar, V., & Sud, K.K. (2006) Image encryption using chaotic logistic map. 926–934. Seyedzade, S. M., Atani, R. E., & Mirzakuchaki, S. (2010). A Novel Image Encryption Algorithm Based on Hash Function. Proceedings of the 6th Iranian Conference on Machine Vision and Image Processing. 10.1109/IranianMVIP.2010.5941167 Sinha, A., & Singh, K. A. (2003). Technique for Image Encryption using Digital Signatures. Optics Communications, 218(4-6), 229–234. doi:10.1016/S0030-4018(03)01261-6 Soutar, C., & Tomko, G. J. (1996). Secure private key generation using a fingerprint. Proceedings of CardTech/SecurTech Conference (Vol. 1, pp. 245–252). Uludag, U., Pankanti, S., & Jain, A. K. (2004). Biometric cryptosystems: issues and challenges. Proceedings of the IEEE, 92(6). Xiao, H. P., & Zhang, G. J. (2006). An Image Encryption Scheme Based On Chaotic Systems. IEEE Proceedings of the Fifth International Conference on Machine Learning and Cybernetics, Dalian (pp. 2707-2711). Younes, A. M. B., & Jantan, A. (2008). Image Encryption Using Block-Based Transformation Algorithm. IAENG International Journal of Computer Science. Zeghid, M., Machhout, M., Khriji, L., Baganne, A., & Tourki, R. (2007) A Modified AES Based Algorithm for Image Encryption. World Academy of Science, Engineering and Technology. Zhou, Y., Bao, L., & Chen, C. L. P. (2014). A new 1D chaotic system for image encryption. Signal Processing, 97, 172–182. doi:10.1016/j.sigpro.2013.10.034

ADDITIONAL READING John, J., & Raj, A. S. (2015). Reliable Biometric Data Encryption Using Chaotic Map. International Journal of Advanced Research in Computer and Communication Engineering, 4(11). Khan, M. K., & Zhang, J. (2007). An intelligent fingerprint-biometric image scrambling scheme. Proceedings of International Conference on Intelligent Computing (Vol. 2, pp. 1141–1151). Kocarev, L. Chaos-based cryptography: a brief overview. Circuits and Systems Magazine, 1(3), 6–21. Ratha, N. K., Connell, J., & Bolle, R. (2001). Enhancing security and privacy in biometrics based authentication systems. IBM Systems Journal, 40(3), 614–634. doi:10.1147/sj.403.0614

47

48

Chapter 4

DNA Sequence Based Cryptographic Solution for Secure Image Transmission Grasha Jacob Rani Anna Government College, India Murugan Annamalai Dr. Ambedkar Government Arts College, India

ABSTRACT With the advent of electronic transactions, images transmitted across the internet must be protected and prevented from unauthorized access. Various encryption schemes have been developed to make information intelligible only to the intended user. This chapter proposes an encryption scheme based on DNA sequences enabling secure transmission of images.

INTRODUCTION Internet has become ubiquitous, faster, and easily manageable by anyone on the earth. Social Networking enables people of all ages and strata to interact with each other in any part of the world through sites like Facebook, Twitter, Linked-In, YouTube, Blogs, Wikis and so on. In today’s information epoch, individuals, businesses, corporations, and countries are interconnected and Information on demand is practically inevitable anytime, anywhere. The essence of global economy is influenced by the internet being available, and it is tough to imagine a day without email or social networking. The idea of being connected anytime, anywhere and having instant information and data sharing certainly comes out with a substantial risk. Security has become more and more of an issue in recent years. Data in transit can be regarded as secure if and only if both the sender and the receiver are capable of protecting the data and the communication between the two hosts is identified, authenticated, authorized and private, meaning that no third party can eavesdrop on the communication between them. Hackers constantly find ways of stealing sensitive data using various techniques. It is estimated that the number of devices connected DOI: 10.4018/978-1-7998-1763-5.ch004

Copyright © 2020, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

to Internet will reach to more than 50 billion around the year 2020.Corporate Espionage has become a reality in this age of the Internet and the global economy. E-commerce transactions are currently plagued with cyber-attacks and are a serious deterrent to the growth of e-commerce globally. In 2008, 4.8 million credit cards were compromised in USA. Revenue losses to the tune of $ 3.3 billion were reported in 2009 from US alone due to cyber-attacks. Successful penetration of Web sites has become the trend of the day. Unfortunately, security issues are more complex no matter how much technology is used. Data Security and Cryptography go hand in hand as cryptography is an accepted and effective way of protecting data. Though cryptography is used commonly in transit, it is now increasingly being used for protecting data at rest as well. Encryption consists of changing the data located in files into unreadable bits of characters unless a key to decode the file is provided. The security of the sensitive information transmitted through an insecure public communication channel poses a great threat by an unintended recipient. Cryptographic techniques help in ensuring the security of such sensitive information. Cryptography enables the sender to securely store or transmit sensitive information across insecure networks so that it can be understood only by the intended recipient. A cryptographic system applies encryption on the information and produces an encrypted output which will be meaningless to an unintended user who has no knowledge of the key. Knowledge of the key is essential for decryption. Defense organizations often use encryption systems to ensure that secret messages will be unreadable if they are intercepted by unintended recipients. Encryption methods can include simple substitution codes, like switching each letter for a corresponding number, or more complex systems that require complicated algorithms for decryption. As long as the coding is kept secret, encryption can be a good method for securing information. On computers systems, there are a number of ways to encrypt images in order to make them more secure. An encryption scheme is unconditionally secure if the ciphertext generated does not contain enough information to determine uniquely the corresponding plaintext no matter how much ciphertext is available or how much computational power the attacker has. With the exception of the one-time pad, no cipher is unconditionally secure. The security of a conditionally secure algorithm depends on the difficulty in reversing the underlying cryptographic problem such as how easy it is to factor large primes. All ciphers other than the one-time pad fall into this category. An encryption scheme is said to be computationally secure ifthe cost of breaking the cipher exceeds the value of the encrypted information the time required to break the cipher exceeds the useful lifetime of the information Shannon introduced two fundamental properties for any cipher to be perfectly secure - diffusion and confusion. The idea of diffusion is to hide the relationship between the cipher text and plain text. Diffusion implies that each bit in the cipher text is dependent on all bits in the plain text i.e., if a single bit in the plain text is changed several or all bits in the cipher text will be changed. The idea of confusion is to hide the relation between the cipher text and the key. This will infuriate the adversary who tries to use the cipher text to find the key. The diffusion effect can be introduced on cipher text by permutation. The confusion effect can be introduced on cipher text by substitution box or S-box. In secure cryptographic schemes, the legitimate user should be able to decipher the messages and the task of decrypting the cipher text should be infeasible for an adversary. But today, the breaking task can be easily performed by a non-deterministic polynomial-time machine. The Data Encryption Standard (DES) is an algorithm with approximately 72 quadrillion possible keys. The security of the DES is based on the difficulty of picking out the right key after the 16-round nonlinear function operations. Boneh et al. describe in detail a library of operations which were useful 49

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

when working with a DNA computer. They estimated that given one arbitrary (plain-text, cipher-text) pair, one could recover the DES key in about 4 months of work. Furthermore, they showed that under chosen plain-text attack it was possible to recover the DES key in one day using some preprocessing. Their method could be generalized to break any cryptosystem which uses keys of length less than 64 bits. This clearly indicates that molecular computing has the ability to break DES. Though cryptography enables in ensuring security to sensitive information, code breakers have come up with various methods to crack the cryptographic systems being developed. Though computational power offered by the revolution in Information Technology paved the way to build new and strong algorithms in cryptography, it is also a strong tool used by cryptanalysts to break the cryptosystems. Hence the subject of finding new and powerful ciphers is always of interest and new directions in cryptography are explored. As traditional cryptographic methods built upon mathematical and theoretical models are vulnerable to attacks, the concept of using DNA computing in the field of cryptography has been identified as a possible technology that brings forward a new hope for unbreakable algorithms. In cryptography, a one-time pad (OTP) is an encryption technique that cannot be cracked when used correctly. In this technique, a plaintext is paired with a random secret key (or pad) using an XOR operation. Then, each bit or character of the plaintext is encrypted by combining it with the corresponding bit or character from the pad using modular addition. If the key is truly random - at least as long as the plaintext, is never reused in whole or in part, and is kept completely secret, then the resulting ciphertext will be impossible to decrypt or break. It has also been proven that any cipher with the perfect secrecy property must use keys with effectively the same requirements as OTP keys. The OTP properties correspond to the characteristics of the unbreakable encryption system defined by Shannon. The XOR operation is reversible. Consider two images – a plaintext image, P and a Key image, K of the same size (dimension) as that of the plain text image. Let E be the resultant image obtained when an XOR operation is performed on P and K. The original image P is obtained (regained or restored) when an XOR operation is once again performed on E and K. E ⇽ P ⊕ K P ⇽ E⊕ K

(1)

The key image used as one time pad must be of the same size as that of the plaintext image to prevent information from being leaked. The primary merit of XOR operation is that it is simple to implement, and that it is computationally inexpensive.

DNA COMPUTING DNA stands for Deoxyribo Nucleic Acid. DNA represents the genetic blueprint of living creatures. DNA contains instructions for assembling cells. Every cell in the human body has a complete set of DNA. DNA is unique for each individual. DNA is a polymer made of monomers called deoxy-ribo nucleotides. Each nucleotide consists of three basic items: deoxyribose sugar, a phosphate group and a nitrogenous base. The nitrogenous bases are of two types: purines (Adenine and Guanine) and pyrimidines (Cytosine and Thymine). The key thing to note about the structure of DNA is its inherent complementarity proposed by Watson and Crick. A binds with T and G binds to C. All DNA computing applications are based on

50

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

Watson-Crick complementarity. DNA computing is an inter disciplinary area concerned with the use of DNA molecules for the implementation of computational processes. The main features of DNA are massive parallelism, intense storage capacity and energy efficiency. Adleman’s pioneering work gave an idea of solving the directed Hamiltonian Path Problem (Travelling Salesman Problem) of size n in O(n) using DNA molecules. The principle used by Adleman lies in coding of information (nodes, edges) in DNA clusters and in the use of enzymes for the simulation of simple calculations. The various operations performed on DNA are synthesized, cutting, ligation, translation, substitution, polymerase chain reaction, detection using gel electrophoresis and affinity purification. Adleman’s work urged other researchers to develop DNA-based logic circuits using a variety of approaches. The resulting circuits performed simple mathematical and logical operations, recognized patterns based on incomplete data and played simple games. Molecular circuits can even detect and respond to a disease signature inside a living cell, opening up the possibility of medical treatments based on man-made molecular software. Lipton extended the work of Adleman and investigated the solution of Satisfiability of Propositional Formula pointing to new opportunities of DNA computing. Research work is being done on DNA Computing either using test tubes (biologically) or simulating the operations of DNA using computers (Pseudo or Virtual DNA computing). In 1997, L. Kari gave an insight of the various biological operations concerning DNA.

DNA Cryptography Taylor (1999) et al. proposed a substitution cipher for plaintext encoding where base triplet was assigned to each letter of the alphabet, numeral and special characters and demonstrated a steganographic approach by hiding secret messages encoded as DNA strands among multitude of random DNA. Decryption was difficult with the use of sub-cloning, sequencing and there was a need of an additional triplet coding table. In 2000, Gehani et al. introduced a trial of DNA based Cryptography and proposed two methods: i) a substitution method using libraries of distinct one time pads, each of which defines a specific, randomly generated, pair-wise mapping and ii) an XOR scheme utilizing molecular computation and indexed random key strings were used for encryption. They used the natural DNA sequences to encode the information and encrypted an image by using the XOR logic operation. Such experiments could be done only in a well-equipped lab using modern technology, and it would involve high cost. Leier et al. (2000) also presented two different cryptographic approaches based on DNA binary strands with the idea that a potential interceptor cannot distinguish between dummies and message strand. The first approach hid information in DNA binary strands and the second designed a molecular checksum. Decryption was done easily using PCR and subsequent gel electrophoresis without the use of sub-cloning, sequencing and additional triplet coding table. Although the approach of generating bit strands shown here had advantages such as rapid readout, it also had practical limitations. One of the limitations was the resolution of the used agarose-gels. Chen (2003) presented a novel DNA-based cryptography technique that took advantage of the massive parallel processing capabilities of biomolecular computation. A library of one time pads in the form of DNA strands was assembled. Then, a modulo-2 addition method was employed for encryption whereby a large number of short message sequences could be encrypted using one time pads. A novel public-key system using DNA was developed by Kazuo et al. (2005) based on the one-way function. The message-encoded DNA hidden in dummies could be restored by PCR amplification, followed by sequencing.

51

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

The YAEADNA algorithm proposed by Sherif et al. (2006) used a search technique in order to locate and return the position of quadruple DNA nucleotide sequence representing the binary octets of plain text characters. Plain text character and a random binary file were given as input and the output PTR was a pointer to the location of the found quadruple DNA nucleotide sequence representing the binary octet. The encryption process was tested on images to show how random the selection of DNA octet’s locations is on the encrypting sequence. Cui et al. (2008) designed an encryption scheme by using the technologies of DNA synthesis, PCR amplification, DNA digital coding and the theory of traditional cryptography. The data was first preprocessed to get completely different ciphertext to prevent attack from a possible word as PCR primers. Then, the DNA digital encoding technique was applied to the ciphertext. After coding sender synthesizes the secret-message DNA sequence which was flanked by forward and reverse PCR primers, each 20mer oligo nucleotides long. Thus, the secret-message DNA sequence was prepared and at last sender generated a certain number of dummies and put the sequence among them. Once the data in encrypted form reached the receiver’s side the reverse procedure was followed to decrypt it. Biological difficult issues and cryptography computing difficulties provided a double security safeguard for the scheme. The intended PCR two primer pairs used as the key of this scheme was designed by the complete cooperation of sender and receiver to increase the security of this encryption scheme. Ning (2009) explained the pseudo encryption methodology based upon the work of Gehani. The plain text was converted to DNA sequences and these sequences were converted to the spliced form of data and protein form of data by cutting the introns according to the specified pattern and it was translated to mRNA form of data and mRNA was converted into protein form of data. The protein form of data was sent through the secure channel. The method did not really use DNA sequences, but only the mechanisms of the DNA function; therefore, the method was a kind of pseudo DNA cryptography methods. The method only simulates the transcription, splicing, and translation process of the central dogma; thus, it was a pseudo DNA cryptography method. Sadeg et al. (2010) proposed a symmetric key block cipher algorithm which included a step that simulated ideas from the processes of transcription (transfer from DNA to mRNA) and translation (from mRNA into amino acids). Though the encryption algorithm (OTP) proposed was theoretically unbreakable, it experienced some disadvantages in its algorithm. These drawbacks had prevented the common use of its scheme in modern cryptosystems. Qinghai (2010) had also proposed a method to protect information, including representing information using biological alphabets to enhance the security of traditional encryption, using DNA primer for secure communication and key distribution, and using the chemical information of DNA bases for steganography. Alice and Bob share a secret DNA sequence codebook. Alice can design a sequence that can maximally match one of the sequences in the codebook and then send the designed sequence to Bob through a public channel. When Bob receives the sequence he would use the non-matching letters in the private sequence as the encryption key. Knowing the public string only, an attacker cannot decrypt the transmitted information. Xuejia et al. (2010) also proposed an asymmetric encryption and signature cryptosystem by combining the technologies of genetic engineering and cryptology. It was an exploratory research of biological cryptology. DNA-PKC uses two pairs of keys for encryption and signature, respectively. Using the public encryption key, everyone can send encrypted message to a specified user, only the owner of the private decryption key can decrypt the ciphertext and recover the message; in the signature scheme, the owner of the private signing key can generate a signature that can be verified by other users with the 52

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

public verification key, but no else can forge the signature. DNA-PKC differs from the conventional cryptology in that the keys and the ciphertext are all biological molecules. The security of DNA-PKC relies on difficult biological problems instead of computational problems; thus DNA-PKC is immune from known attacks, especially the quantum computing based attacks. In 2010, in the image encryption algorithm based on DNA sequence addition operation combined with logistic chaotic map to scramble the location and value of pixel of an image presented by Qiang et al., a DNA sequence matrix was obtained by encoding the original image and it was divided into some equal blocks and two logistic maps. DNA complementarity and DNA sequence addition operations were utilized to add these blocks. DNA sequence matrix was decoded to get the encrypted image. The experimental results and security analysis showed that the proposed algorithm had larger key space and resisted exhaustive, statistical and differential attacks. Qiang et al. (2012) presented a novel image encryption algorithm based on DNA subsequence operations that uses the idea of DNA subsequence operations (such as elongation operation, truncation operation, deletion operation, etc.) combining with the logistic chaotic map to scramble the location and the value of pixel points from the image. The experimental results and security analysis showed that the proposed algorithm was easy to be implemented, had good encryption effect and a wide secret key’s space, strong sensitivity to secret key, and had the abilities of resisting exhaustive attack and statistic attack but the defect was its weak ability of resisting differential attack.

DNA Coding An electronic computer needs only two digits, 0 and 1 for coding information. As a single strand of DNA is similar to a string consisting of a combination of four different symbols, A, C, G and T, DNA coding should reflect the biological characteristics of the four nucleotide bases- A, C, G and T along with the Watson-Crick complementary rule (A is complementary to T and C is complementary to G) [46]. Out of the twenty four combinations of the four nucleotides, only eight combinations given within parenthesis satisfy the complementary rule of the nucleotides. (00011011 - C T A G, 00011011 - C A T G, 00011011 - G T A C, 00011011 - G A T C, 00011011- T C G A, 00011011- T G C A, 00011011 - A C G T, 00011011 - A G C T). Figure 1 represents the molecular structure of the four nucleotides. In accordance with the increasing molecular weight of the four nucleotides, (C -111.1 g/mol, T - 126.1133 g/mol, A - 135.13 g/mol and G - 151.13 g/mol)C T A G, is the best coding pattern and is used as DNA Coding. According to DNA Coding Technology, C denotes the binary value 00, T denotes 01, A denotes 10 and G denotes 11 so that Watson-Crick complementary also holds good. Table 1 gives the DNA Coding. This pattern could perfectly reflect the biological characteristics of the four nucleotide bases and has biological significance. Figure 1. a) Cytosine b) Thyminec) Adenine d) Guanine

53

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

Table 1. DNA coding Digital Value

DNA Base

Molecular Weight g/mol

00

C

111.1

01

T

126.11

10

A

135.13

11

G

151.13

Axiomatic Definition of DNA Algebra DNA algebra is an algebraic structure defined on a set of elements B{C, T, A, G} together with two binary operators ‘V’ and ‘^’ provided the following Huntington postulates are satisfied. 1a. Closure with respect to the operator V. 1b. Closure with respect to the operator ^. 2a. An identity element with respect to V, designated by C: x V C = x, ∀x∈B

(2)

2b. An identity element with respect to ^, designated by G: x ^ G = x, ∀x∈B

(3)

3a. Commutative with respect to V. x V C = C V x, ∀x∈B

(4)

3b. Commutative with respect to ^ . x ^ G = G ^ x, ∀x∈B

(5)

4a. V is Distributive over ^. x V (y ^z) = (x V y) ^ (x V z), ∀x,y,z∈B

(6)

4b. V is Distributive over ^. x ^ (y V z) = (x ^ y) V (x ^ z), ∀x,y,z∈B For every element x∈B, there exists an element x’ ∈B (called the complement of x) such that x V x’ = G and

54

(7)

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

x ^ x’ = C

(8)

There exists atleast two elements x, y ∈B, such that x ≠ y. The following De’ Morgan’s laws also hold good. (x V y) ’ = x’ ^ y’, ∀x,y∈B

(9)

(x ^ y) ’ = x’ V y’, ∀x,y∈B

(10)

The primitive logic operations OR, AND, NOT and XOR can be carried out and the results will be obtained as given in the characteristic tables Table 2, Table 3, Table 4 and Table 5.

DNA Sequence Based Image Representation The term image refers to a two-dimensional light intensity function, denoted by f(x,y), where the value of f at spatial coordinates(x,y) gives the intensity(brightness) of the image at that point. As light is a form of energy f(x,y) must be nonzero and finite, that is f(x) must lie between zero and infinity (0 < f(x,y) < ∞). A digital image is an image f(x,y) that has been discretized both in spatial coordinates and brightness. A digital image can be considered a matrix (two dimensional array) whose row and column indices identify a point in the image and the corresponding matrix element value identifies the gray level at that point. The elements of such a picture array are called pixels. Each pixel of the image consists of 8 bits. Using DNA coding principle, substituting C for 00, A for 01, T for 10 and G for 11, each pixel of the DNA image is represented as a quadruple nucleotide sequence. Table 2. AND characteristic table >

Table 4. XOR characteristic table

C

T

A

G

⊕

C

T

A

G

C

C

C

C

C

C

C

T

A

G

T

C

T

C

T

T

T

C

G

A

A

C

C

A

A

A

A

G

C

T

G

C

T

A

G

G

G

A

T

C

Table 3. OR characteristic table V

C

T

Table 5. NOT characteristic table A

G

C

C

T

A

G

T

T

T

G

G

A

A

G

A

G

G

G

G

G

G

X

C

T

A

G

~X

G

A

T

C

55

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

Arithmetic and Logic Operations on Images Arithmetic and logic operations between pixels are used extensively in most of the branches of image processing and are generally carried out on images pixel by pixel. The principle use of image addition is for image averaging - to reduce noise. Image addition is mainly done using XOR operation for brightening the image and for image security applications. Image subtraction is a basic tool in medical imaging where it is used to remove static background information. Image multiplication or division is used to correct gray level shading resulting from non-uniformities in illumination or in the sensor used to acquire the image. Arithmetic operations can be done “in place” such that the result of performing an operation can be stored in that location in one of the existing images. With DNA computing, two DNA images can be added using parallel addition of the rows of the two images.

Biological Operations on DNA Sequences The following biological operations can be performed on DNA sequences in a test tube to program the DNA computer.

Synthesis In standard solid phase DNA synthesis, a desired DNA molecule is built up nucleotide by nucleotide on a support particle in sequential coupling steps. For example, the first nucleotide (monomer), say A, is bound to a glass support. A solution containing C is poured in, and the A reacts with the C to form a two-nucleotide (2-mer) chain AC. After washing the excess C solution away, one could have the C from the chain AC coupled with T to form a 3-mer chain (still attached to the surface) and so on.

Separation Separation is the process of separating the strands by length using gel electrophoresis.

Merging Merging or mixing is the process of pouring the contents of two test tubes into a third one to achieve union. Mixing can be performed by rehydrating the tube contents (if not already in solution) and then combining the fluids together into a new tube, by pouring and pumping.

Extraction Extraction is the process of extracting those strands that contain a given pattern as a substring by using affinity purification.

Annealing Annealing is the process of bonding together two single-stranded complementary DNA sequences by cooling the solution. Annealing in vitro is also known as hybridization. 56

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

Melting Melting is the process of breaking apart a double-stranded DNA into its single-stranded complementary components by heating the solution. Melting in vitro is also known as denaturation.

Amplification Amplifying is making copies of DNA strands by using the Polymerase Chain Reaction. PCR is an in vitro method that relies on DNA polymerase to quickly amplify specific DNA sequences in a solution. PCR involves a repetitive series of temperature cycles, with each cycle comprising three stages: denaturation of the guiding template DNA to separate its strands, then cooling to allow annealing to the template of the primer oligonucleotides, which are specifically designed to flank the region of DNA of interest and finally, extension of the primers by DNA polymerase. Each cycle of the reaction doubles the number of target DNA molecules, the reaction giving thus an exponential growth of their number.

Cutting Cutting is the process of cutting DNA double-strands at specific sites by using restriction enzymes.

Ligation Ligation is the process of pasting DNA strands with complementary (compatible) sticky ends using ligase.

Substitute Substitute is the process of substituting - inserting or deleting (cutting and ligation) DNA sequences by using PCR site-specific oligonucleotide mutagenesis.

Marking/ Unmarking Single strands are marked by hybridization, that is, complementary sequences are attached to the strands, making them double-stranded. The reverse operation is unmarking of the double-strands by denaturing, that is, by detaching the complementary strands. The marked sequences will be double-stranded while the unmarked ones will be single-stranded.

Destroying The marked strands are destroyed by using exonucleasesor by cutting all the marked strands with a restriction enzyme and removing all the intact strands by gel electrophoresis.

Detecting and Reading Given the contents of a tube, say “yes” if it contains at least one DNA strand, and “no” otherwise. PCR may be used to amplify the result and then a process called sequencing is used to actually read the solution. 57

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

Operations and Functions Used The operations of molecular biology and functions that are used for this work can be equivalently formalized as follows:

Synthesis Creation of DNA sequences for the image data is referred to as Synthesis. DNA sequences are made up of four bases – A, C, G and T. According to the DNA Coding Technology, C denotes 00, A – 01, T - 10 and G – 11.Each pixel of eight bits is converted into a quadruple nucleotide sequence. 10010011 → TACG

(11)

Translation When the positions of sequences are translated, the sequences are interchanged. Translation is represented as P1P2P3P4 ↔P5P6P7P8

(12)

Substitution Each quadruple nucleotide sequence is substituted by the value returned by the DNA Sequence Crypt function. Substitution is represented by the following expression. V ←DNASequenceCryptfn(P1P2P3P4)

(13)

Detect Detect searches for a quadruple nucleotide sequence of the image starting from a random position in the DNA sequence file and returns true if a match is found and false otherwise.

Re-Substitution Each value, V in the encrypted image is replaced by the corresponding quadruple nucleotide sequence from that position in the DNA Sequence File. P1P2P3P4← V

(14)

Re-Synthesis Re-synthesis is the process of converting each sequence into its digital form. TACG → 10010011

58

(15)

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

DNA Sequence Crypt Function DNA Sequence Crypt function is a function that returns one of the many positions of the quadruple DNA sequence in the key DNA sequence file. A one to many DNA Sequence Crypt function is a one-to-many function d(x), which has the following three properties: 1. A pointer, h maps an input quadruple nucleotide sequence, x to one of the many positions obtained in random in the key DNA sequence file. 2. Ease of computation: Given d and an input x, d(x) is easy to compute. 3. Resistance to guess: In order to meet the requirements of a cryptographic scheme, the property of resistance to guess is required of a crypt function with input x, x1 and outputs y, y1. As similar quadruple nucleotide sequence that occur in a plain text are mapped to different positions in the DNA nucleotide sequence file(one to many mapping), it is difficult for a recipient to guess the plain-text.

DNA Sequence Database DNA sequences can be obtained from GenBank®, the NIH genetic sequence database, an annotated collection of all publicly available DNA sequences. GenBank is part of the International Nucleotide Sequence Database Collaboration, which comprises the DNA DataBank of Japan (DDBJ), the European Molecular Biology Laboratory (EMBL), and GenBank at NCBI. These three organizations exchange data on a daily basis. Query sequence(s) to be used for a BLAST search should be pasted in the ‘Search’ text area. It accepts a number of different types of input and automatically determines the format or the input. A sequence in FASTA format begins with a single-line description, followed by lines of sequence data. The description line (defline) is distinguished from the sequence data by a greater-than (“>”) symbol at the beginning. It is recommended that all lines of text be shorter than 80 characters in length. In cryptographic applications, long DNA sequences are generally used.

HYBRID ENCRYPTION SCHEME Hybrid encryption scheme proposed is a symmetric encryption scheme. Symmetric encryption algorithms use an identical secret key for encryption and decryption process and the key is sent to the receiver through a secure communication channel. Ke = Kd = K

(16)

The requirement of a symmetric algorithm is that both the sender and the receiver know the secret key, so that they can encrypt and decrypt the information easily.

59

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

The proposed hybrid encryption scheme is a combination of a cryptosystem using XOR-OTP and substitution proposed by Gehani et al. and the DNA-based Implementation of YAEA Encryption Algorithm proposed by Sherif et al given in Figure 2.

Figure 2. The DNA YAEA Encryption Algorithm proposed by Sherif et al.

60

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

Principle In the hybrid encryption scheme, the image to be encrypted and the key image are synthesized - transformed into DNA images and XOR One-Time Padding is performed using the substitution operation (a complex operation which is a combination of cutting and ligating operations) and an intermediate image is obtained. Both cutting and ligation use the same enzymes that organisms use for the maintenance of their own DNA. Cutting uses a restriction endonuclease. Table 4 represents the truth table for the XOR operation on DNA nucleotides. The intermediate image is scanned for four nucleotides at a time and one of the many positions (index) of the same quadruple DNA nucleotides sequences as that of the scanned nucleotides of the image in the gene sequence file is detected. The index is stored as a double dimensional array and the resultant encrypted image (2D array) is sent to the receiver. The key image used for OTP is sent to the receiver through a secure communication channel. The sender and receiver should also agree upon the same DNA sequence file that is used for encryption and hence decryption. Since this is a symmetric key encryption scheme, the same DNA sequence file is used during the decryption process. In the decryption process, the received image (2D array) is actually pointers (index) to the DNA sequence file. The indices are substituted by quadruple nucleotide sequences starting from that index and the DNA image thus obtained is XOR-ed with the key DNA image (synthesized- converted into DNA image) and the resultant decrypted image is the original image that was transmitted in the encrypted form. The encryption and decryption algorithms of the hybrid encryption scheme are pictorially represented in Figure 3 and Figure 4. ALGORITHM 1. HYBRID_CRYPT Input: X [image file] to be encrypted, Y [image file] –key image, R[Binary file that contains DNA nucleotides sequence], Output: Encrypted image E 1. SYNTHESIS a. Convert image file X into its DNA sequence X ← DNA [X] b. Convert key image file Y into its DNA sequence Y ← DNA [Y] 2. SUBSTITUTION // OTP X ← X⊕ Y 3. DETECTION and SUBSTITUTION For each quadruple DNA nucleotide sequence in X search from a random location RND (Z) in a binary file are represented in the form of a single strand DNA sequence. If the correct pattern is found, its location I is then recorded If the search is successful Then store I in E; Else Repeat step 3 End Algorithm

61

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

ALGORITHM 2. HYBRID_DECRYPT Input: Encrypted image E, Y [image file] – key image, R [Binary file that contains DNA nucleotides sequence] Output: Decrypted image X 1. SYNTHESIS Convert E into DNA sequence 2. DETECTION and RE-SUBSTITUTION F ← DNA sequence represented by E [I] in the binary file R 3. SUBSTITUTION and RE-SYNTHESIS F ← F ⊕ Y // O TP Convert F into its binary equivalent and display the image X. End Algorithm.

Experimental Results Matlab R2008a was used to simulate the DNA operations on a MiTAC Notebook PC with Intel® Core™ 2Duo CPU T6400 @ 2.00 GHz, 2 GB RAM, 32 bit operating system. Experiments were performed using different images of different sizes to prove the validity of the proposed algorithm. Figure 5 shows an example of an original image (plaintext), key image, encrypted image (Ciphertext) and the decrypted image (Figures 3-5).

Figure 3. Encryption process

62

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

Figure 4. Decryption process

Figure 5. a) Original image of size 128 x 128b) Key image of size 128 x 128 c) encrypted image d) decrypted image

Cryptanalysis A good information security system should be able to protect confidential images and should be robust against statistical, cryptanalytic and brute-force attacks. The level of security that the hybrid encryption algorithm offers is its strength. The proposed method is examined through statistical analysis, sensitivity to key changes and key space analysis.

Statistical Analysis The encrypted image should not have any statistical similarity with the original image to prevent the leakage of information. The stability of the proposed method is examined via statistical attacks - the histogram and correlation between adjacent pixels. Histogram Analysis An image histogram is a graphical representation of the number of pixels in an image as a function of their intensity and describes how the image-pixels are distributed by plotting the number of pixels at each intensity level.

63

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

Figure 6. a) Original image (Elaine) b) histogram of original image c) encrypted image d)histogram of encrypted image e)original image (Elaine) f)histogram of original image g)encrypted image h) histogram of encrypted image

An attacker will find it difficult to extract the pixels statistical nature of the original image from the encrypted image if the histograms of the original image and encrypted image are different. Figure 6 a) - h) show the original Elaine.bmp of size 128 x 128, its histogram, the encrypted image, its histogram and the original Lena.bmp of size 128 x 128, its histogram, the encrypted image and its histogram. The histograms of the encrypted images (Figure 6 d) and h)) are fairly uniform and significantly different from the histograms of the original images (Figure 6 a) and e)). It is clearly observed that the histograms of the encrypted images look quite similar (though the original images are different) and are completely different from that of the histograms of the original images and do not provide any information regarding the distribution of gray values to the attacker; Hence the proposed algorithm can resist any type of histogram based attacks and this strengthens the security of encrypted images significantly. Correlation Coefficient Analysis In most of the plaintext-images, there exists high correlation among adjacent pixels, while there is a little correlation between neighboring pixels in the encrypted image. It is the main task of an efficient image encryption algorithm to eliminate the correlation of pixels. Two highly uncorrelated sequences have approximately zero correlation coefficient. The Pearson’s Correlation Coefficient is determined using the formula: γ =

n ∑xy − (∑x )(∑y ) n(∑x )(∑x ) 2

2

n(∑y )(∑y ) 2

2

(17)

where x and y are the gray-scale values of two adjacent pixels in the image and n is the total number of pixels selected from the image for the calculation. Table 6 tabulates the correlation coefficient calculated for the original and encrypted images. If there is no linear correlation or a weak linear correlation, γ is close to 0. A value near zero means that there is a random, nonlinear relationship between the two adjacent pixels and a value near one indicates that there is a linear relationship between adjacent pixels. It is

64

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

Table 6. Correlation coefficient Correlation Coefficient Image

Horizontal

Vertical

Diagonal

Original

Encrypted

Original

Encrypted

Original

Encrypted

Elaine.bmp

0.93076

-0.00801

0.94215

0.00918

0.88403

-0.02338

Lena.bmp

0.88875

0.00024

0.95104

0.00724

0.86972

0.01820

Camera.bmp

0.90886

-0.03015

0.94411

-0.01280

0.88152

-0.01162

Figure 7. Correlation coefficient relationships

clear from Figure 7 that there is negligible correlation between the two adjacent pixels in the encrypted image. However, the two adjacent pixels in the original image are highly correlated as the correlation coefficient value is close to 1.

Differential Attacks Attackers often make a slight change for the original image, use the proposed algorithm to encrypt the original image before and after changing, and compare two encrypted images to find out the relationship between the original image and encrypted image. This is known as differential attack. Known-Plaintext and Chosen Plaintext Attacks For encryption with a higher level of security, the security against both known-plaintext and chosenplaintext attacks are necessary. Chosen/Known-plain text attacks are such attacks in which one can access/choose a set of plain texts and observe the corresponding encrypted texts. Figure 8a) is the mask image obtained by the XOR operation of a Plain image and its encrypted image. Figure 8b) shows an unsuccessful chosen/known-plain text attack using the proposed algorithm. Algorithm Test Input: C,Z - Plain Images, C1, Z1 – Cipher image Output T - Boolean Step 1:M← C ⊕ C1// Mask Image 65

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

Figure 8. a) XOR Mask b) Failed attack to crack the encrypted image

Step 2: If Z = M ⊕Z1 //Z1 - encrypted image of the plain image Z then T ⟵ True//Unknown encrypted image is decrypted else T ⟵ False //resists Chosen/ Known Plaint Text attack end if End Algorithm

Brute Force Attack A Brute Force Attack or exhaustive key search is a strategy that can be used against any encrypted data by an attacker who is unable to take advantage of any weakness in an encryption system that would otherwise make his task easier. It involves systematically checking all possible keys until the correct key is found. The key to this encryption algorithm is an image which is of the same size as that of the image to be encrypted. Moreover, the aspect of bio-molecular environment is more difficult to access as it is extremely difficult to recover the DNA digital code without knowing the correct coding technology used. When an intruder gets the encrypted image and tries to decrypt the encrypted image without knowing the correct DNA digital coding technology, it would not be decrypted at all. An incorrect coding will cause biological pollution, which would lead to a corrupted image. Key Sensitivity Analysis The key sensitivity is an essential feature for any good cryptographic algorithm to guarantee the security of the proposed system against the brute-force attack. The key sensitivity of a proposed method can be observed in two different ways: i) the cipher image cannot be decrypted correctly if there is a small difference between the encryption and decryption keys and ii) if a different sequence file is used for encryption and decryption. Figure 9 and Figure 10 clearly show that the decryption depends on the key image and DNA sequence file used for encryption. A different key image or a different DNA sequence file will not decrypt the encrypted image.

66

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

Figure 9. Decryption when a different sequence is used

Figure 10. Decryption when a different key image is used

CONCLUSION DNA based encryption is the beneficial supplement to the existing mathematical encryption. DNA binary strands support feasibility and applicability of DNA-based Cryptography. The security and the performance of the DNA based cryptographic algorithms are satisfactory for multi-level security applications of today’s network. The proposed hybrid encryption scheme using DNA technology can resist brute-force, statistical and differential attack.

REFERENCES Adleman, L. (1994). Molecular Computation of Solutions to Combinatorial Problems. Science, 266(5187), 1021–1024. doi:10.1126cience.7973651 PMID:7973651 Amos, M., Paun, G., Rozenberg, G., & Salomaa, A. (2002). Topics in the theory of DNA computing. Theoretical Computer Science, 287(1), 3–38. doi:10.1016/S0304-3975(02)00134-2 Chen, J. (2003). A DNA-based biomolecular cryptography design. IEEE International Symposium on Circuits and Systems, (pp. 822–825). Cui, G. Z., Qin, L. M., Wang, Y. F., & Zhang, X. (2007). Information Security Technology Based on DNA Computing. IEEE International Workshop on Anti-counterfeiting, Security, Identification. 10.1109/ IWASID.2007.373746 Cui, G. Z., Qin, L. M., Wang, Y. F., & Zhang, X. (2008). An Encryption Scheme using DNA Technology. International Conference on Bio-Inspired Computing: Theories and Applications. Forouzan, B. A., & Mukhopadhyay, D. (2007). Cryptography and Network Security. Tata: McGraw Hill.

67

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

Gao, Q. (2010). Biological Alphabets and DNA-based Cryptography. American Society for Engineering Education. Garzon, M. H., & Deaton, R. J. (2000). Biomolecular computing and programming: A definition. Kunstliche Intelligenz, 1, 63–72. Gehani, A., LaBean, T., & Reif, J. (2000). DNA Based Cryptography. DIMACS Series in Discrete Mathematics and Theoretical Computer Science., 54, 233–249. Gifford, D. K., & Winfree, E. (2000). DNA Based Computers V. American Mathematical Society. Gothelf, K. V., & LaBean, T. H. (2005). DNA-programmed assembly of nanostructures. Organic & Biomolecular Chemistry, 3(22), 4023–4037. doi:10.1039/b510551j PMID:16267576 Gupta, V., Parthasarathy, S., & Zaki, M. J. (1997). Arithmetic and logic operations with DNA. Proc. 3rd DIMACS Workshop on DNA-based Computers. Heider, D., & Barnekow, A. (2008). DNA-based watermarks using the DNA Crypt algorithm. BMC Bioinformatics, 8(1), 176. doi:10.1186/1471-2105-8-176 PMID:17535434 Kahn, D. (1967). The Code breakers. New York: McMillan. Kari, Daley, Gloor, Siromoney, & Landweber. (1999). How to Compute with DNA. Foundations of Software Technology and Theoretical Computer Science, 269-282. Kari, L. (1997). From Micro-soft to Bio-soft. Biocomputing and Emergent Computation Publications, World Scientific, 26, 146–164. Leier, A., Richter, C., Banzhaf, W., & Rauhe, H. (2000). Cryptography with DNA binary strands. Bio Systems, 57(1), 13–22. doi:10.1016/S0303-2647(00)00083-6 PMID:10963862 Li, J., Zhang, Q., Li, R., & Shihua, Z. (2008). Optimization of DNA Encoding Based on Combinatorial Constraints. ICIC Express Letters ICIC International, 2, 81–88. Ning, K. (2009). A pseudo DNA cryptography Method. CoRR. Retrieved from http://arxiv.org/ abs/0903.2693 Norcen, R., Podesser, M., Pommer, A., Schmidt, H. P., & Uhl, A. (2003). Confidential storage and transmission of medical image data. Computers in Biology and Medicine, 33(3), 273–292. doi:10.1016/ S0010-4825(02)00094-X PMID:12726806 Qiang, Z., Xue, X., & Wei, X. (2012). A Novel Image Encryption Algorithm based on DNA Subsequence Operation. The Scientific World Journal. Reif, J. (1997). Local Parallel Biomolecular Computation. 3rd DIMACS workshop on DNA based computers. Richa, H. R., & Phulpagar, B. D. (2013). Review on Multi-Cloud DNA Encryption Model for Cloud Security. Int. Journal of Engineering Research and Applications, 3(6), 1625–1628. Richter, C., Leier, A., Banzhaf, W., & Rauhe, H. (2000). Private and Public Key DNA steganography. 6th DIMACS Workshop on DNA Based Computers.

68

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

Risca, V. I. (2001). DNA-based Steganography. Cryptologia, Taylor and Francis, 25(1), 37–49. doi:10.1080/0161-110191889761 Rothemund. (1996). A DNA and restriction enzyme implementation of Turing machines. DNA Based Computers, 6, 75-120. Rothemund, P. W. K., Papadakis, N., & Winfree, E. (2004). Algorithmic self-assembly of DNA Sierpinski triangles. PLoS Biology, 2(12), e424. doi:10.1371/journal.pbio.0020424 PMID:15583715 Rozenberg, G., Bäck, T., & Kok, J. (2012). Handbook of Natural Computing. Springer. doi:10.1007/9783-540-92910-9 Rozenberg, G., & Salomaa, A. (2006). DNA computing: New ideas and paradigms. Lecture Notes in Computer Science, 7, 188-200. Sadeg, S. (2010). An Encryption algorithm inspired from DNA. IEEE International Conference on Machine and Web Intelligence. Sakakibara, Y. (2005). Development of a bacteria computer: From in silico finite automata to in vitro and invivo. Bulletin of EATCS, 87, 165–178. Sánchez, R., Grau, R., & Morgado, E. (2006). A Novel Lie Algebra of the Genetic Code over the Galois Field of Four DNA Bases. Mathematical Biosciences, 202(1), 156–174. doi:10.1016/j.mbs.2006.03.017 PMID:16780898 Shannon, C. E. (1949). Communication theory of secrecy system. Journal of Bell System Technology, 28(4), 656–715. doi:10.1002/j.1538-7305.1949.tb00928.x Sherif, T. A., Magdy, S., & El-Gindi, S. (2006). A DNA based Implementation of YAEA Encryption Algorithm. IASTED International Conference on Computational Intelligence. Siromoney, R., & Bireswar, D. (2003). DNA algorithm for breaking a propositional logic based cryptosystem. Bulletin of the European Association for Theoretical Computer Science, 79, 170–177. Stallings. (n.d.). Cryptography and Network Security – Principles and Practice (5th ed.). Prentice Hall. Susan, K. L., & Hellman, M. E. (1994). Differential-Linear Cryptanalysis. LNCS, 839, 17–25. Tausif, A., Sanchita, P., & Kumar, S. (2014). Message Transmission Based on DNA Cryptography [Review]. International Journal of Bio-Science and Bio-Technology, 6(5), 215–222. doi:10.14257/ ijbsbt.2014.6.5.22 Taylor, C., Risca, V., & Bancroft, C. (1999). Hiding messages in DNA Microdots. Nature, 399(6736), 533–534. doi:10.1038/21092 PMID:10376592 Terec. (2011). DNA Security using Symmetric and Asymmetric Cryptography. International Journal of New Computer Architectures and their Applications, 34-51.

69

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

Udo, Saghafi, Wolfgang, & Rauhe. (2002). DNA Sequence Generator: A program for the construction of DNA sequences. DNA Computing, 23-32. Vazirani, V. (2004). Approximation Algorithms. Berlin: Springer. Volos, C. K., Kyprianidis, I., & Stouboulos, I. (2013). Image encryption process based on chaotic synchronization phenomena. Signal Processing, 93(5), 1328–1340. doi:10.1016/j.sigpro.2012.11.008 Wang, X., & Wang, Q. (2014). A novel Image Encryption Algorithm based on Dynamic S-Boxes constructed by chaos. Nonlinear Dynamics, 75(3), 567–576. doi:10.100711071-013-1086-2 Wang, Z., & Yu, Z. (2011). Index-based symmetric DNA encryption algorithm. Fourth International Congress on Image and Signal Processing. Watson, J. D., & Crick, F. H. C. (1953). Molecular structure of nucleic acids: A structure for De-oxy ribose nucleic acid. Nature, 25(4356), 737–738. doi:10.1038/171737a0 PMID:13054692 Weichang, C., & Zhihua, C. (2000). Digital Coding of the Genetic Codons and DNA Sequences in High Dimension Space. ACTA Biophysica Sinica, 16(4), 760–768. Westlund, B. H. (2002). NIST reports measurable success of Advanced Encryption Standard. Journal of Research of the National Institute of Standards and Technology. Winfree. (1980). The Geometry of Biological Time. Academic Press. Winfree, E. (1996). On the Computational Power of DNA Annealing and Ligation. 1st DIMACS workshop on DNA based computers. Wu, Y., Noonan, J. P., & Agaian, S. (2011). NPCR and UACI Randomness Tests for Image Encryption. Journal of Selected Areas in Telecommunications, 4, 31–38. Wu, Y., Zhou, Y., Noonan, J. P., & Agaian, S. (2013). Design of Image Cipher Using Latin Squares. Information Sciences, 00, 1–30. Xiao, G., Lu, M., Qin, L., & Lai, X. (2006). New field of cryptography: DNA cryptography. Chinese Science Bulletin, 51, 1139–1144. Xiao, J. H., Zhang, X. Y., & Xu, J. (2012). A membrane evolutionary algorithm for DNA sequence design in DNA computing. Chinese Science Bulletin, 57(2), 698–706. doi:10.100711434-011-4928-7 XueJia. (2010). Asymmetric Encryption and signature method with DNA technology. Information Sciences, 53, 506–514. Youssef, Emam, & Saafan, & AbdElghany. (2013). Secured Image Encryption Scheme Using both Residue Number System and DNA Sequence. The Online Journal on Electronics and Electrical Engineering, 6(3), 656-664. Yunpeng, Z., Fu, B., & Zhang, X. (2012). DNA cryptography based on DNA Fragment assembly. IEEE International Conference Information Science and Digital Content Technology.

70

DNA Sequence Based Cryptographic Solution for Secure Image Transmission

Zhang, M., Sabharwal, L., Tao, W., Tarn, T.-J., Xi, N., & Li, G. (2004). Interactive DNA sequence and structure design for DNA Nano applications. IEEE Transactions on Nanobioscience, 3(4), 286–292. doi:10.1109/TNB.2004.837918 PMID:15631140 Zhang, X. C. (2008). Breaking the NTRU public key cryptosystem using self-assembly of DNA tilings. Chinese Journal of Computers, 12, 2129–2137.

This research was previously published in the Handbook of Research on Modern Cryptographic Solutions for Computer and Cyber Security edited by Brij Gupta, Dharma P. Agrawal, and Shingo Yamaguchi; pages 505-527, copyright year 2016 by Information Science Reference (an imprint of IGI Global).

71

72

Chapter 5

IPHDBCM:

Inspired Pseudo Hybrid DNA Based Cryptographic Mechanism to Prevent Against Collabrative Black Hole Attack in Wireless Ad hoc Networks Erukala Suresh Babu K.L. University, India C. Nagaraju Y.V. University, India M.H.M. Krishna Prasad J.N.T. University, India

ABSTRACT Secure communication is one of the basic requirements for any network standard. Particularly, cryptographic algorithms have gained more popularity to protect the communication in a hostile environment. As the critical information that is being transferred over the wireless adhoc networks can be easily acquired and is vulnerable to many security attacks. However, several security communication threats had been detected and defended using conventional symmetric and asymmetric cryptographic mechanism, which are too difficult and resource consuming for such mobile adhoc networks. Recently, one of the severe security threats that have to be detected and defend in any type of network topology is blackhole attack and cooperative blackhole. Because of its severity, the black hole attack has attracted a great deal of attention in the research community. Comprehensively the results of the existing system conclude that the black hole attack on various mobile adhoc networks is hard to detect and easy to implement. This paper addresses to detect and defend the blackhole attack and cooperative blackhole attack using hybrid DNAbased cryptography (HDC) mechanism. Moreover, the proposed method upsurge the security issue with the underlying AODV routing protocol. Eventually, This Hybrid DNA-based Cryptography (HDC) is one of the high potential candidates for advanced wireless ad hoc networks, which require less communication bandwidth and memory in comparison with other cryptographic systems. The simulation results of this proposed method provide better security and network performances as compared to existing schemes. DOI: 10.4018/978-1-7998-1763-5.ch005

Copyright © 2020, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

IPHDBCM

1. INTRODUCTION There is a necessity to design and develop a secure wireless mobile ad hoc network (SWMANETs), particularly useful for battlefield applications in order to perform security-sensitive operations. Unlike wireless network with fixed infrastructure that makes use of access points to communicate, a MANET is an infrastructure less network does not require any centralized administration. however, network elements of these networks are needed to be deployed rapidly with reasonably low cost. One of the primary concerns of these networks requires resilient security service, which are more vulnerable to limited physical insecurity of mobile nodes, as these nodes are disposed to attacks. These attacks are performed in both reactive and proactive routing protocols, which can roughly be relegated into two major categories such as active and passive attacks (Konate, K., 2011; Gopi, A. P et al., 2015 ; Kumar, S. A et.al 2015). The malicious nodes pretend to be as a trusted router by advertising the spurious service requests to disrupt the normal routing operation and to deny the services to authorized nodes, which leads to a DOS attack. In active attack, the malicious router originates the attack by modifying the information in the network. The black hole attack is one such type of active assaults that can be performed against both reactive and proactive routing protocols. To be more specific, each black hole node impersonates the source and destination node by sending an imitated path request to the destination node and imitated path reply to the source node that was taking place in route discovery phase to claim that, it has the optimal route information. Finally, the black hole node consumes the packet, and simply drops the packets, that reduce the network performance as shown in figure-1. On the other hand, under the context of information and network security domain, it is necessary to provide an unbreakable cryptosystem to protect the data that we transmit over the network; Open Systems Interconnection (OSI) security architecture provided a systematic security solution for different layers of networks. In the routing layer of the OSI model, it is essential to design a secure protocol that can defend the black hole attack and cooperative black hole attack against on-demand routing protocol (Osathanunkul, 2011; Dasgupta, 2012). More importantly, the security services such as authentication, data confidentiality, nonrepudiation and data integrity services must be incorporated into these on-demand routing protocols. The following are some of the challenging issues to secure against black hole attack and cooperative black hole attack. The first challenging issue is to secure the routing protocols against the black hole attack. This problem has not properly addressed in most of the existing secure routing protocols or if addressed, there are very expensive in terms of bandwidth and limited computational capabilities. The second challenging issue is to defend the black hole attack against adhoc routing protocols that dynamically changes the topology, (i.e., what kind of key management and authentication schemes are needed? Unlike of Wireless networks, MANET cannot use any certificate authority (CA) server).The third challenging issue is the existing secure routing protocols may not efficient or feasible to scale, as these protocols produce heavy traffic load and requires intensive computations. This paper mainly addresses all the above issues using hybrid DNA based cryptographic mechanism to defend and detect a black hole attack and cooperative black hole attack against AODV routing protocol. We call this protocol, as Secure Routing Protocol using Hybrid DNA-based Cryptography (SRP-HDC) that establishes cryptographically secure communication links among the communicating mobile nodes. The rest of the paper is as follows. Section-II specifies the related work that had proposed a large class of ad-hoc routing protocols and attacks against MANET in the literature. Section-III enumerates the hybrid cryptosystem that prevents the collabrative black hole attack against on-demand routing protocols by authentication and encryption mechanism. Section-IV presents the detection mechanism 73

IPHDBCM

Figure 1. Illustrates the black hole scenario

against collabrative black hole attack which will be incorporated into the secure routing route discovery procedures. Section-V specifies the security analysis and the simulation results for better evaluation of our theoretical work using network simulator NS-2. Finally, conclusion with future work are presented in Section-VI

2. RELATED WORK The proposed SRPHDC protocol for mobile adhoc network may fall broadly into two categories, first, integrating the Inspired Pseudo Biotic DNA predicated Cryptographic approach into the existing AODV routing protocol and second, detecting the black hole attack against AODV in MANETs. This section will discuss the revolution of the Black hole problem, which is a genuine security issue in mobile adhoc network that affect its performance. In this problem, a malevolent node makes use of the weakness of the AODV routing protocol to advertise itself having the optimal path to the destination with the aim of intercepting the packets. This malevolent node will drop all the packets or drops selectively to accomplish a denial-of-service (DOS) attack. Recently, many proposals had been proposed in defending, avoiding and detecting black hole nodes in mobile adhoc networks(Woungang, 2012 ; Lu, 2009; Bindra, 2012). In this paper, we collect and introduce the mechanisms that are proposed in recent years. In (Tsou et al., 2011). introduced a reverse tracing technique to detect and prevent the black hole nodes against DSR routing protocol. In (Baadache, A & Belmehdi A, 2010), Baadache et al. proposed a novel method based on the principle of Merkle tree to detect the black hole nodes against adhoc networks. However, their method experiences more computational overhead on routing. In (Jain et al., 2010). proposed a mechanism for detecting the cooperative malicious nodes based neighborhood monitoring of data blocks between source and destination. In (Anita et al., 2010) proposed a certificate based authentication method for recognizing the black hole attacks in mobile ad hoc networks. In (Lu, S et al., 2009) proposed a Secure AODV routing protocol against black hole attack for MANETS by showing the security limitations of AODV. In (Raj et al.,2009) proposed a strategy to identify and separate the

74

IPHDBCM

blackhole attacks in mobile ad hoc networks. In their methodology, the algorithm experiences extreme overhead due to the handling of ALARM control messages and the threshold value should be updated at every time interval. In (Weerasinghe et al., 2007) introduces a system to prevent the cooperative blackhole attacks in mobile ad hoc networks. However, in their solution, black hole attack can be prevented; the extra control packets will cause more overhead in the network and high latency in the network. In (Deng et al., 2002) projected a mechanism to keep away the black hole nodes in mobile adhoc networks. However, this technique suffers from a cooperative black hole attack. (Shurman et al.,2004) proposed two different techniques to detect the black hole attack against MANETS. However, their solution cannot handle to detect multiple black hole attack. (Marti et al., 2000) proposed the misbehavior detection using the watchdog and the path rater. However, their solution cannot handle partial dropping of packets. (Tissieres et al., 2003) proposed the confidant protocol, which, monitor the neighbor nodes that makes use of reputation system. (Kozma et al.,2009) proposed a detection mechanism using the REACT system. However, in their solutions only gives the information of transmitting packets, but no information about forwarding path. (Woungang et al.,2012) introduced a mechanism to identify a black hole attack against the DSR routing protocol. However, in their solutions cannot handle cooperative black hole attacks and computation overhead is present. Most of the solutions discussed above are used to detect or avoid black hole attacks on reactive routing protocols in the mobile ad hoc networks. However, most of the methods are used to either detect or preventing the black hole attacks. In this paper, a novel method is presented based on the AODV protocol in which the adversaries are detected based upon the close neighborhood of the range and avoiding the black hole using multiple path between source and destination and finally defending by integrating inspired pseudo biotic DNA based cryptographic mechanisms to the existing AODV routing protocol.

3. SECURE ROUTING AGAINST AODV PROTOCOL In the layer-5 of the OSI network model, most of the possible attacks will be either on routing information or data tampering. The prevention of these attacks against on-demand routing protocols can be performed using authentication and encryption mechanism.

3.1. Overview of AODV Routing Protocol: In brief, to summarize the AODV(Perkins et al., 2001; Babu, E. S et al., 2013) routing protocol. AODV is the predominant on-demand routing protocol that offers low processing, low network utilization, ability to adapt the dynamic conditions and low memory overhead. We used this AODV as an underlying protocol to protect from the black hole attack. The functionality of AODV is usually initiated with the route discovery process, whenever a valid route is not present, and another mechanism is route maintenance, as AODV fails to maintain lifelong route between the sending node and receiving node, due to the high mobility by nature. During the route discovery process, if the originator needs of a route, broadcasts route request (RREQ) packet (with regular information and security related information) to its neighboring nodes, which is described in the next section-III(C). Once the neighbor node obtains a RREQ message from the originator, it broadcast the same RREQ message to its next hop with its current route. This process will be continual until it acquires the actual destination.

75

IPHDBCM

Figure 2. Illustrates route discovery process of modified AODV

This proposed work modifies the original AODV protocol. The slight modification is done at the destination side. To be more specific, the destination node broadcast the Route Reply (RREP) packet back to all its neighbor nodes with the current route until reaches to the source node to create multiple route, instead of unicasting as in the original AODV routing protocol, as shown figure-2. To adapt the dynamic topology environment in the route discovery process of AODV in adhoc networks, we used multiple, possibly disjoint, routes/path (Kim et al., 2006; E. S et al., 2013) between source and destination. This modified AODV routing protocol has an ability to discover responds by unicasting the RREP packet to the source node. Instead of unicasting with single RREP packet, it broadcasts RREPs Reply (RREP) packet back to all its neighbor nodes with the current route until reaches to the source node. Once the first RREP message received by the source node, then it can begin sending the data to the destination, late arrived RREPs will be reserved or saved for future purpose. The alternative multiple paths between source and destination can be used for two purpose, first if the primary path fails to send the packets to the destination. Second, after detection of black hole node (both single and cooperative black hole node), source node will diverted the traffic with alternate route to the destination. Finally, once the path is entrenched, the nearest neighbor nodes will monitor the link status for the active routes. The nodes that do not conformed the neighbor rating based on neighbor profile will be eliminated from the route as described from Section-D(i,ii).

3.2. Background of DNA Cryptography: In order to understand the rudimentary principles of DNA Cryptography (Gehani at al., 2003 ; Leier et al., 2000; Ning et al 2009) in a emerge area of DNA Computing, it is necessary to address the background details of central dogma of molecular biology, that is, how a DNA sequence is actually transcript and translated into a protein sequence as shown in figure-3. DNA (Deoxyribo Nucleic Acid) is the funda-

76

IPHDBCM

mental hereditary material that stores genetic information found in almost every living organisms ranging from very small viruses to complex human beings. It is constituted by nucleotides which forms polymer chains. These chains are also known as DNA strands. Each DNA nucleotides contains a single base and usually consists of four bases, specifically, Adenine (A), Guanine (G), Cytosine(C), and Thymine (T) represent genetic code. These bases reads from the start promoter which forms the structure of DNA strand by forming two strands of hydrogen bonds, one is A with T and another is C with G; These DNA sequences are eventually transcript and interpreted into chains of amino acids, which constitutes proteins.

3.3. Node Authentication using Hybrid DNA based Cryptosystem: This section describes the hybrid DNA based cryptosystem (Babu et al., 2015), which is used to verify the data integrity and authenticate the mobile nodes. Moreover, this hybrid approach makes use of both the public and private key-based schemes. • •

First, In order to authenticate the mobile nodes, public-key cryptography is used with a key pair of DNA based public keys, DNA based private keys, and it is used to establish the session keys. Second, pseudo DNA based symmetric cryptography is used to verify the data integrity. This can be achieved with the private key, which is shared by the two nodes.

In other words, symmetric encryption will be used to achieve integrity and confidentiality, while asymmetric encryption will provide to authenticate the members of mobile nodes. Subsequently, the above method can be succeed with the following assumption. • • • •

First, clearly, according the characteristics of MANETS, initial trustees must exist among the mobile nodes. Second, pairwise DNA based mutual secret keys must exist between the nodes. Third, during the initialization phase of the network, we embed the unique ID, an initial key pair of DNA based private key, and DNA based public key for every node. Finally, the shared secret keys are used to reboot from PKI, which can distribute and generate a key pair of DNA based private key and DNA based public key for every node.

Next, In order to authenticate the mobile nodes and to establish the session keys, the following framework is used. • • •

Initially, a node ‘S’ generates a pair of DNA secret key K(PR,S) and public key K(PU,S) and distributed the public key K(PU,S) to the node ‘D’ by using Public Key Infrastructure or Trusted Authority. Similarly, a node ‘D’ also generates a pair of DNA private key K(PR,D) and public key K(PU,D) and distributed the DNA public key K(PU,D) to the node ‘S’ by using PKI or CA. If the node ‘S’ and node ‘D’ are 1-hop neighbors then Node ‘S’ can authenticate the node ‘D’ by issuing a signed certificate with its DNA private key. Here certificate is a proof of node ‘D’ ID and DNA public key with ‘S’ signature.

77

IPHDBCM

Figure 3. Central dogma of molecular biology

• •

i.

If any of the intermediate nodes ‘w’ want the read the signed certificate, he/she should hold ‘S’ DNA public key and then node ‘w’ can read and trust that node by bind it along its DNA public key. Finally, N-hop one to one intermediate nodes can quickly create a DNA private key by using threeway handshake based on the key information and availability of certificates in the PKI (Capkun, S et al., 2003) Secure Route Discovery Process on AODV routing protocol:

Whenever, source node wants to send the data to the destination. It initiates the path discovery process only, when no valid route is present to the destination. Consequently, it broadcasts Route Request (RREQ) packet by creating pairwise DNA private key/shared key (the procedure of symmetric DNA based cryptography is discussed in section III-C(iii) with neighbor nodes, until RREQ reaches to the destination node. To achieve this, first, source node generates pseudo random number and signs that number to create the certificate with its DNA secret key using asymmetric cryptosystem, subsequently the RREQ packet is secured by a Message Digest (MD5) algorithm (Babu, E. S., 2015, September ; Babu, E. S., (2015).), finally, the signature and generated hash value is attached to the RREQ control packet, and forwards to its intermediate nodes. Second, the intermediate node having source node DNA public key will verify the signed certificate and then decrypt the message that contains the shared secret key that can be summarized as.

(

) ((

)

)

cmq ,i + h cmq ,i + K (sk , j ) + E E K (sk , j ) ,K (i, pub) K(i, pri )

(1)

Where cmq is RREQ control message that contains original message (M), identity Number (IN) of a node that forwards the original message and sequence number (SeqNo) of the message, which can be written in (2). h(cmq,i+K(sk,j)) represents the keyed hash Message algorithm with a shared key K(sk,s) on message cmq,i which can be written in (3). Finally ‘+’ denotes the concatenation of strings; the suffix

78

IPHDBCM

‘i’ is the number of intermediate nodes form source to destination i=1,2,3,4,…,n and suffix ‘j’ is the number of keys created the source node ‘s’ which is usually represented as j=k1,k2,k3,…,kn .

(

)

h cmq ,i + K (sk , j )

(2)

h(cmq,i+K(sk,j))

(3)

Once RREQ control packet reaches with the valid route to the destination, then it verifies the signature and decrypts the shared key with its private key and reply with a RREP control packet that transmits back to the source which can be summarized as

K sk , s

(4)

ii. To Authenticate the Communications between nodes using Pseudo DNA Asymmetric Cryptography: Definition of Public Key Cryptography: An Asymmetric Cryptosystem is mathematical function with three-tuple i = 1, 2, 3, 4, ….n where: ε is the encryption algorithm and j =k1, k2 , k 3 , ……., kn is the Decryp-

tion algorithm and cmq ,i =M +(IN f ) + SeqNo is a randomized key generation algorithm. It returns two keys in which one is a DNA public key denoted by K(Dpub,s) and another DNA private key denoted by

(

)

K(Dpri,s); we write a1←K(Dpub,s) and a2←K(Dpri,s); h cmq ,i + K(sk , j ) is a stateful public key algorithm. The deterministic encryption algorithm takes a plaintext ‘p’ along with the random key ‘a1’ and returns

(

) ((

)

)

a cipher text ‘q’; we denote as q←Ea(p); cm p,i + h cm p,i + K(sk , j ) + E E K(sk , j ) ,K(i, pub) K(i, pri) is a decryption algorithm which takes the cipher text ‘q’ along with DNA private key ‘a2’ and returns the equivalent plaintext ‘p’ ; we denote as p←Da(q) where p∈{0,1,A,G,T,C}*: finally we represent the stateful public key algorithm with Da2(Ea1(p))=p for all p∈{0,1,A,G,T,C}*. Suppose, a source node ‘s’ want to send the data to the destination. It initiates with path discovery process with one of the 1-hop intermediate node ‘r’, then it generates the random key (Here, the Key will number of the splices, the starting code of the frame and removed length of the pattern codes) which will provided as shared DNA private key between source ‘s’ and destination ‘r’. and then source node encrypts the DNA shared secret key ‘sk’ by utilizing its neighbor DNA public key K(pub,r) and then once again source node encrypts the encrypted DNA based shared secret key ‘sk’ by utilizing its own private key K(pub,s). The output of encrypted shared key ‘sk’ provides signature for the RREQ control packet and secured by a Message Digest (MD5) algorithm. Once the RREQ control packet reaches with the valid route to the destination, then it verifies the integrity of message (signature) and decrypts the shared key with its private key and reply with a RREP control packet that transmits back to the source node. Hence, source node ‘s’ can authenticate all intermediate node up to the destination node by creating the shared key ‘sk’ and distribute to all the nodes.

79

IPHDBCM

Figure 4. shows the architectural diagram of asymmetric cryptography used as a part of SRP-HDC work

iii. Communications model for exchanging the shared key using Pseudo DNA based symmetric cryptography: In (Babu, E. S et al., 2016; Nagaraju, C et al., 2016) We proposed a novel method called pseudo symmetric DNA based cryptographic mechanisms. Particularly, whenever a node needs to transmit the packet, first, the node should share the unique DNA based secret key with the source. i.e. private key are shared by the two participating nodes. Moreover, to protect the control or data traffic, the source node ‘s’ can simple use the DNA based shared key and sent to the destination ‘d’. The complete format is shown below: M + h(M + SKsd)

(5)

where SKsd is the shared key which is the part of the message M that is shared between source ‘s’ and destination ‘d’ .Here pseudo DNA based symmetric cryptography is mainly used to achieve data integrity and confidentiality. To perform the above communications model, the following steps can be described briefly as shown in figure-5. 1. Alice takes the plaintext (Secret Key) and it converts into binary form which in turn converts into DNA form 2. Alice will scan DNA form of information to generate the variable length random key by generating the No of the splices from the specified DNA pattern, the starting code of the DNA frame to find out the introns, introns places and removed length of the pattern codes i,e. introns are removed from the specified DNA sequence as the first round of Key Generation. 3. With the help of random key (splicing system), Alice will transcript the DNA sequence into the mRNA strand

80

IPHDBCM

4. After Generating mRNA Strand, Alice also generate the variable length random sub key by generating the No of the splices from the specified mRNA pattern, the starting code of the mRNA frame, introns places and removed length of the pattern codes as the second round of processing 5. Again, the spliced mRNA strand are translated into the amino acids, which forms protein sequence, as shown in Algorithm-5 (Hint: The mapping of codons to amino acids is done with the help of genetic code table. Usually table consists of 61 codons, which are mapped 20 amino acids) 6. Next, the protein sequence (Cipher Text) will be sent to the Bob through public channel. 7. The random variable length key is comprised of number of splices, the starting index, pattern codes length of the introns, the positions and places of the introns, the cut out the introns, random mapping of codon-amino acids will form the symmetric key to decrypt the cipher text (protein sequence), which is also sent to the Bob through a secure channel 8. On Bob’s (Receiver) side, when he receives random keys and protein form (Cipher text) of data from Alice through the secure channel, then he can perform the decryption process. 9. Bob decrypts the cipher text message using the random key reversible translation to recover mRNA sequence from protein sequence, and then recover DNA form of information, in the reverse order as Alice encrypt the information. 10. Bob can then recover then binary form of information, and finally gets what Alice sent him.

Figure 5. Pseudo DNA symmetric cryptosystem

81

IPHDBCM

4. DETECTION OF BLACK HOLE ATTACK AGAINST AODV PROTOCOL: This sub-section performs the next task, after developing the secure route discovery process on AODV routing protocol, the detection scheme against black hole attack should be incorporated into secure route discovery procedures. To detect the black hole attack, we assume and summarize the following. 1. Every node has an exclusive unique identifier and a pair of DNA based public/secret keys. 2. Initially a compromised participating node will be in routing until it is detected. 3. The source node will establish the DNA shared key with the destination via with all its intermediate nodes by utilizing the hybrid DNA key management scheme as discussed in Section-III(C). 4. There are sufficient uncompromised intermediate nodes in the arbitrary network. so that packets can reach with different routes to the destination. a. Detection of Single Black hole attack: Initially, the nodes are allowed to build up with the trust, based on neighbor profile and behaviors of the nodes. Specifically, neighborhood profile includes all the features such as number of data packets or RREQ control packets sent as well as received, number of ACK or RREP packets are sent or forwarded and received from/to the neighbors, number of RREQ/RREP/data packets dropped. To detect the misbehaving nodes, first, the neighbor profile approach is used. In this approach, the nearest node can identify the misbehaving nodes by monitoring the network traffic of its neighboring nodes. Here, a profile is used to detect abnormal behavior in the network. Second, the nearest node makes use of the responses, such as (No response, Shutdown and Blacklisting) to detect the misbehavior of nodes. In our work, first two cases (No response, shut down response) of responses were chosen, as these two cases are more suitable for ad-hoc networking. The nearest node waits for the response from the neighbor nodes, the neighbor node communicates back to the nearest node in the form of response (i.e. nearest neighbor will response back (shut down or no responses) to the source node or intermediate nodes). To achieve this k-Nearest Neighbor algorithm (describe below) is used to compute the k nearest neighbor by calculating the distance between the nodes. To be more specific, let us assume that source node ‘s’ sends the protected message to the destination node ‘d’ via intermediate node ‘M’. We also assume that even though the intermediate node ‘M’ is initially authenticated, it can be compromised, during the route discovery phase. According to the authentication, malicious node ‘M’ could not tamper the packets that was received from source, but simply drop the packets instead of forwarding the packets to the destination. Meanwhile one of the nearest node ‘NN’ of the intermediate node ‘M’ response with shut down or no responses packet back to the source node or intermediate nodes. Based upon the response received and neighbor profile, the source node assume that intermediate node ‘M’ is misbehaving, however, the source node sends the packet through different path or route to the destination. Moreover, a Destination node also able to detect intermediate node ‘M’ is misbehaving. Finally, the source node gives the neighbor rating (described below) based on neighbor profile.

Nearest Neighbor (NN) Problem: It is an optimization problem (Andoni A, 2009) that achieves the distance to the requested neighbor. To shorten this problem, we convey that a node ‘x’ is a K-nearest neighbor of a node ‘y’ i.e. the radius between node ‘x’ and node ‘y’ is at most K. This NN algorithm, either gives back K-nearest neighbor 82

IPHDBCM

Figure 6. An illustration of a K-near neighbor query

node or no such nodes exists. The requested nearest neighbor of the node ‘y’ is the node ‘x’. However, both the nodes n1 and n2 are K-nearest neighbors of ‘y’ as illustrated in figure-6.

Neighbors Rating Neighbor rating is usually calculate with the Beta distribution function.The probability distributions function of any two events is more often represented with Beta function, which is defined as ∏ = (ED , ,K)

(6)

where 0≤e≤1, α>0, β>0 Let us assume the formation of nodes with two possibilities , the probability of ‘e’ is estimated by Beta distribution function based on past observations of D, and it is resresented with α and β α= 1 + observed value of e β= 1 + observed value of Here, the binary process is normally corresponds like a node’s behavior, the number of positive events ‘e’ and number of negative events a1 ← K(Dpub,s) are measured over a period of time accordingly. To calculate the neighbor rating, it is necessary to assign flexible weights to different events; for instance, more importance will be given for data packet dropping rathen than control packet dropping. “Case-1: if the direct neighbor rating from the node ‘j’ to node ‘i’ is of 1-hop neighbor, then direct neighbor rating (DR) is calculated based on its observations which is defined as: a2 ← K(Dpri,s)

(7)

83

IPHDBCM

Pi,j = latest positive variable actions weights of node j is observed by node i ni,j= latest negative variable actions weights of node j is observed by node i The direct rating can be represented as, E (M, K)

(8)

where (C, K)

where γ is the past behavior which is part of direct neighbor rating; 0≤γ≤1 In this work, we had added more variable weights to monitor the current behavior rather than the past behavior, which is denoted as small ‘γ’, because the malicious node (blackhole attack) can misuse the positive rating by limiting the weights. However, the past history will be deleted, as the time progresses to make effective rating. Finally, the entire computation is based upon the neighbor profile history. Moreover, this technique gives the efficient results in high mobility network.” The total rating is usually denoted as TRi,j that gathers all the direct neighbor rating DRi,j for the set of one-hop neighbors N, which is represented as DRn,j for every n∈N. Here, either N is the neighbors set that can trusted, or their information on rating will be sent for the abnormality test. The abnormality test of a node i checks the behavior based on given rating value DRn,j, if the difference of DRn,j is large than the expected rating value TRi,j then it is reject otherwise it will accept. This condition makes more robust detection method against black hole attack. The abnormality test is defined as: p ← Da (q )

(9)

After detecting the malicious node based on neighbor direct rating, the source node will diverted the traffic with different route to the destination. Hence, intrusion of malicious node effect to network becomes weaker. Therefore, we can conclude that more paths reduce malicious node intrusion to network. ii. Detection of Cooperative Blackhole attack: The below figure-7 depicts the cooperative black hole attack (Babu, E. S et al., 2016 ; Ramaswamy, S et al., 2003 ; Su, M. et al., 2010 ; Swarna, M. et al, 2016). which is more challenging attack to detect. In our design, source node has transfer its shared secret key with all its k-hop-neighbors using key management schemes. Suppose, source node ‘s’ sends the protected message to the destination node ‘d’ via intermediate node ‘M1 and M2’. We also assume that even though the intermediate node ‘M1 and M2’ is initially authenticated, it can be compromised, during the route discovery phase. According to the authentication, malicious node ‘M1’ could not tamper or drop the message that was received from source, nevertheless, it may forward the message to M2 node. In this case, M2 is also comprised and may drop the message instead of forwarding the message to the destination or intermediate node. To detect

84

IPHDBCM

Figure 7. Illustration of an indirect foreign neighbor pair (nj, nk)

the cooperative misbehaving nodes, first, the nearest foreign neighbor profile approach is used. In this approach, each node finds its closest nearest foreign neighbor (2-hop-neighbor) by establishing DNA shared keys between the sender and neighbor nodes. To be more specific, Let nj be the nearest node to M1 and nk be nearest neighbor node to M2 and it is nearest foreign neighbor to s that estimate the history by neighbor rating of M2 with the help of nk. This setup is depicted below: The node nj is the neighbor to M2 but cannot reach to M2, and node nk completely know the profile of M2 since nk is already neighbor to nj, who gives opinion (misbehaving) about M2. Subsequently, In a future behavior between M1 and M2 will be monitored by both nj and nk, the probability of this nodes behavior will be rated by both the nodes nj and nk is: m1 and m2 is monitored by nk Xm1,m2(y) = indicator variable for node nj for the confirmation of malicious node m2 at encounter y p {0, 1, A, G, T, C} ; *

(10)

Let Ds,m1(n) all the encounters (neighbor rating) between S and m1 is monitored by nj Xs,m1(n) = indicator variable for node s for the confirmation of malicious node m1 at encounter y

(

)

Da 2 Ea1 (p) = p for all

(11)

The interpretation of the equations 10th and 11th which gives the probability of nj and nk, that agree and approves the nodes M1 and M2 is misbehaving. Hence, the nearest foreign neighbor can identify the misbehaving nodes by monitoring the network traffic of its neighboring nodes. After detecting the 85

IPHDBCM

malicious node based on foreign neighbor indirect rating, the source node will diverted the traffic with different route to the destination. Eventually, intrusion of malicious node effect to network becomes weaker. Once again, we can conclude that more paths reduce malicious node intrusion to the network. In summary, the single black hole and cooperative black hole intrusion can be identified without the need of expensive signatures, as these signatures, which can be used to defend the route from end to end.

5. SECURITY ANALYSIS In this section, we investigate the security analysis of SRPHDC (Secure Routing Protocol using Hybrid DNA based Cryptography) by evaluating its strength in the presence of a black hole attack and cooperative black hole attack against AODV routing protocol which is described in Section-D(i,ii)5.3.3. Moreover, we also compared a SRPHDC routing protocol with ARAN (Sanzgiri, K et al., 2005.) protocols. It is observed that SRPHDC is more secure than ARAN. Additionally, this approach effectively reduces the processing power requirements in comparison of existing secure routing protocols, which might take excessive processing power for different kinds of mobile ad hoc networks. As discussed earlier, SRPHDC assumes public key management sub-system, which is capable of securely verifying the identity and public key for a given node. Furthermore, SRPHDC provides the basis of security services such as authentication, confidentiality, and nonrepudiation and data integrity. Finally, we deliberate, how this proposed method satisfies the basic requirements in comparison of several existing (including ARAN) secure routing protocol. We also considered general routing attacks such as Blackhole attacks, participation of unauthorized nodes, Securing Shortest Paths, fabricated routing messages, Spoofed Route Signaling Attack parameters etc. Figure 8. Throughput of AODV Vs SRPHDC

86

IPHDBCM

Figure 9. Throughput of modified AODV Vs SRPHDC

Figure 10. Throughput of AODV Vs SRPHDC in the presence of black hole nodes

•

Black Hole Attacks: In SRPHDC, a node is monitored by its nearest neighbor nodes, which calculates the neighbor rating, based on neighbor profile. If the node is misbehaving, the nearest neighbor nodes will identify and notify to the source node, then the source node will send the packets with alternate routes with different DNA public keys. In particular, to formally characterize the security offered by SRPHDC, we proposed secure route discovery process which is successful with hybrid biotic cryptosystem based on DNA based cryptography which makes use of both the symmetric key (to achieve the data integrity) and asymmetric key-based schemes (to

87

IPHDBCM

Figure 11. Throughput of modified AODV Vs SRPHDC in the presence of black hole nodes

•

•

•

•

88

authenticate the mobile nodes and to establish the session keys). Moreover, DNA based cryptography appears as a rapidly growing field which provides ultra-compact information storage, vast parallelism, and exceptional energy efficiency that minimize the major overhead produced by the major existing cryptosystem. On the other hand, ARAN fails to defend the cooperative blackhole attack and incurs more overhead. Unauthorized Participation: In SRPHDC, authorized nodes are performed during route discovery and computation process. Particularly, only authorized nodes have both private and public key (asymmetric or shared key). During the route discovery process, the source node encrypts the control packets, which is authenticated by each intermediate node. Hence, unauthorized nodes will not generate any fallacious control packets and prohibited to join a route. Fabricated Routing Messages: In SRPHDC, all the authorized intermediate nodes will verify the routing message that was signed by the sending node, as routing packet is secured through an encryption key. Moreover, Symmetric DNA based cryptography is used to secure the session key, which make more difficult for intruder to fabricate the routing message, on the other side ARAN fails to prevent such type of attack. Spoofed Route Signaling: The source node will issue the encrypted route discovery packets by signing the certificate with its private key. Only either intermediate node or destination node, who have the right key can respond to the route discovery process and decrypt them. Hence, the spoofing node cannot launch any type of attack. This prevents completely the impersonation attacks. Securing Shortest Paths: In SRPHDC, during the route discovery process, the source and destination node can built multiple paths. The destination node can choose a secure minimum path among different route from the source as described in Section-III(A). Moreover, This SRPHDC scheme is designed in such a way, only the authenticated immediate neighbors can access the routing packets, which ensures that an intruder cannot inject false routing packets into secure route discovery process. While ARAN does not guarantee a shortest path, but offers a quickest path to the destination.

IPHDBCM

Figure 12. Routing load of AODV Vs SRPHDC in the presence of black hole nodes

Figure 13. Routing load of modified AODV Vs SRPHDC in the presence of black hole nodes

6. SIMULATION RESULTS AND PERFORMANCE ANALYSIS To study the feasibility of our theoretical work, we have implemented and evaluated the Secure Routing Protocol Using Hybrid Biotic DNA based Cryptography method using network simulator [NS2],which is a software program running in Ubuntu-13.04 and conducted a series of experiments to evaluate its effectiveness. The experiment results show that this method is more efficient and increase the power against black hole attacks.

89

IPHDBCM

Figure 14. Packet delivery ratio of AODV Vs SRPHDC in the presence of black hole nodes

Figure 15. Packet delivery ratio of AODV, SRPHDC, ARAN

The experimental values were obtained by evaluating the SRPHDC with the current version of the AODV routing protocol, by including the behavior of malicious (black hole) nodes into the simulations. Moreover, to improve the performance of SRPHDC, we had modified AODV routing protocol with the following assumptions are made on the misbehaving node. 1. The source node will establish the session keys on demand, during the route-discovery process of AODV routing protocol.

90

IPHDBCM

Table 1. NS-2 simulation parameters of SRPDHC NS-2 Parameters Propagation model No of Nodes Transmission Range Simulation Time Simulation Area Node Mobility Traffic Type Data Payload Size Node Pause Time Maximum node speed Key Size

Two Ray Ground 10, 25, 50,75,100 250m 500 Seconds 750m X 750m Model Random Waypoint FTP/TCP 512 Bytes/Packet 0-20s 1-20m/s 1024 bits Key

2. Once the route is established between the source and destination, then one of the intermediate nodes acts as a black hole node that drops both the data and routing packets. Our simulations are mainly used to compare between SRPHDC with AODV routing protocol with and without the presence of malicious nodes. Moreover, we also compare SRPHDC with modified AODV with and without the presence of malicious nodes and ARAN respectively. To evaluate the SRPHDC, we considered various performance metrics. • • • •

Packet Delivery Ratio (PDR): It is the ratio between the number of application layer data packets that are received correctly by all the destination node, and the number of application layer data packets originated by the source node. Average End-to-End-Delay: The Average delay is usually measured in all the correctly received packets. The average time taken to deliver a data packet from the source node and the time taken to be received a data packet at the destination. Routing Overhead: The total number of application layer data packets that have been received by the destination node at a given simulation time t. Total Overhead: The total number of control (non-data) packets that have been delivered by the nodes in the network at the simulation t.

In our first scenario, the experimental values were obtained by evaluating the SRPHDC with the current version of the AODV routing protocol, by including the behavior of malicious (black hole) nodes into the simulations. Here, SRPHDC is compared to original AODV routing protocol because the implementation of SRPHDC is based on AODV. The following Table-1 gives the simulation parameters, which are used to compare SRPHDC with AODV, SRPHDC with modified AODV and ARAN respectively. Particularly, In order to calculate the performance of this proposed work, we had collect the simulated data by running the simulation up to 500 second with an input of 10, 25, 50, 75, 100 nodes. Specifically Figure-8 depicts the total throughput of SRPHDC in comparative to AODV without presence of black hole nodes with various simulating nodes by 10,25,50,75,100 respectively. It is observed that, if the number of misbehaving nodes is very less, then AODV continues delivering the packets correctly to the destination without the need of SRPHDC, because, in this case packet dropping is not serious between source and destination. 91

IPHDBCM

The figure-9 shows total throughput of SRPHDC and modified AODV without presence of black hole nodes with various simulating nodes by 10,25,50,75,100 respectively. In this case, it gives more throughput with modified AODV and SRPHDC compared to normal AODV routing protocol. Figure-10 depicts the throughput of SRPHDC and AODV routing protocol in the presence of 2, 4, 6, 8, 10 black hole nodes out of 10, 25, 50, 75, and 100 nodes respectively. It is observed that AODV drastically decreases the throughput in the presence of 10% to 40% black hole nodes, the black hole will be part of the route in AODV that result in a low packet delivery ratio (PDR) or high packet drop. While, SRPHDC still delivers the data packets with almost the same amount of data as depicted in the figure-8 and figure-9. Figure-11 depicts the throughput of SRPHDC and modified AODV routing protocol in the presence of 2, 4, 6, 8, 10 black hole nodes out of 10, 25, 50, 75, 100 nodes respectively. Indeed, it is observed that modified AODV will drop the packets less than normal AODV in the presence of 10% to 40% black hole nodes. Whereas SRPHDC still delivers the data packets with almost the same amount of data as normal AODV and modified AODV routing protocol. Moreover, SRPHDC tries to detect the black hole nodes and forwards the packets in alternate path that assures successful packet delivery. The figure-12 and figure-13 depicts the total overhead of SRPHDC in comparison with AODV and SRPHDC in comparison with modified AODV respectively. Here, it is observed that SRPHDC gives slight overhead than AODV due to the security mechanism and less overhead than modified AODV in the presence of different numbers of blackhole nodes. The reason is that SRPHDC can detect the blackhole nodes based on neighbor rating and remove them from routing. Moreover, SRPHDC gives less overhead than ARAN, because SRPHDC make use of DNA computing which provides parallel computation and efficient storage capacity . As shown in figure-14, the packet delivery fraction of SRPHDC is higher in all scenarios (10, 25, 50, and 75,100) and almost identical in AODV routing protocol at different speeds. However, AODV drastically decreases in the presence of malicious node, which delivers fewer data packets and routing packets to the destination.

Figure 16. End-to-end delay of AODV Vs SRPHDC in the presence of black hold nodes

92

IPHDBCM

Figure 17. End-to-end delay of SRPHDC and ARAN

While, SRPHDC detects the black hole nodes and removed from the routing based on neighbor rating. Hence, SRPHDC using modified AODV is extremely efficient in the route discovering, monitoring the routing process and maintaining multiple path for delivery of data packets in the presence of malicious node. Furthermore, it is also seen that SRPHDC always outperforms normal AODV. The figure-15 depicts the packet delivery ratio of ARAN and SRPHDC with different pause times. While, SRPHDC has a higher PDR than ARAN. Since, it delivers more data packets than ARAN to the destination As shown in figure-16. End to End Delay of SRPHDC and AODV at different speeds. AODV reduces the PDR in high mobility (maximum speed increases from 1.5 and 10 milliseconds) conditions, because more link breakage between source and destination, Therefore, as maximum speed increases the PDR decreases. As compared to AODV, SRPHDC always has a lower End-to-End packet latency in high mobility conditions. Moreover, SRPHDC make use of multipath routing, if one path fails, an alternate path will immediately be available. The figure-17 shows the End-to-End packet latency of SRPHDC and ARAN. In this case, it is observed that SRPHDC has lower latency than ARAN, Since, SRPDHC make use of DNA computing which provides parallel computation that causes less overhead of cryptographic operations than ARAN. In ARAN, while processing the control packets, each and every node has to verify the signature and replace with its own digital signature that cause more overhead using conventional cryptography and additional delays at each node. Therefore, the End-to-End packet latency increases.

7. CONCLUSION In this paper, we have proposed a scheme for defending and detecting the blackhole attack against modified AODV routing protocol in mobile ad-hoc wireless networks. This SRPHDC protocol which establishes cryptographically secure communication among the nodes. Hybrid DNA-based Cryptography (HDC) is one of the high potential candidates for advanced wireless ad hoc networks which require less

93

IPHDBCM

computational power, communication bandwidth and memory in comparison with other cryptographic systems. The simulation results of this work provide better security, less computation overhead and network performances as compared to existing AODV and ARAN schemes. As a conclusion, DNA cryptography is an emerge new idea and very promising field where research can be done in great innovation and development. but these DNA cryptography lacks the related theory which is nevertheless still an open problem to model the good DNA cryptographic schemes.

REFERENCES Al-Shurman, M., Yoo, S. M., & Park, S. (2004, April). Black hole attack in mobile ad hoc networks. Proceedings of the 42nd annual Southeast regional conference (pp. 96-97). ACM. 10.1145/986537.986560 Andoni, A. (2009). Nearest neighbor search: the old, the new, and the impossible [Doctoral dissertation]. Massachusetts Institute of Technology. Baadache, A., & Belmehdi, A. (2010). Avoiding black hole and cooperative black hole attacks in wireless ad hoc networks. arXiv preprint arXiv:1002.1681. Babu, E. S., Nagaraju, C., & Prasad, M. H. M. (2015, September). A Secure Routing Protocol against Heterogeneous Attacks in Wireless Adhoc Networks. Proceedings of the Sixth International Conference on Computer and Communication Technology 2015 (pp. 339-344). ACM. Babu, E. S., Nagaraju, C., & Prasad, M. K. (2013). An Implementation and Performance Evaluation Study of AODV, MAODV, RAODV. Mobile Ad hoc Networks, 4, 691–695. Babu, E. S., Nagaraju, C., & Prasad, M. K. (2013). An Implementation and Performance Evaluation of Passive DoS Attack on AODV Routing Protocol in Mobile Ad hoc Networks. International Journal of Emerging Trends & Technology in Computer Science, 2(4), 124–129. Babu, E. S., Nagaraju, C., & Prasad, M. K. (2015). Analysis of Secure Routing Protocol for Wireless Adhoc Networks Using Efficient DNA Based Cryptographic Mechanism. Procedia Computer Science, 70, 341–347. doi:10.1016/j.procs.2015.10.029 Babu, E. S., Nagaraju, C., & Prasad, M. K. (2016). Efficient DNA-Based Cryptographic Mechanism to Defend and Detect Blackhole Attack in MANETs. Proceedings of International Conference on ICT for Sustainable Development (pp. 695-706). Springer Singapore. 10.1007/978-981-10-0129-1_72 Babu, E. S., Nagaraju, C., & Prasad, M. K. (2016). Inspired Pseudo Biotic DNA based Cryptographic Mechanism against Adaptive Cryptographic Attacks. International Journal of Network Security, 18(2), 291–303. Babu, E. S., Nagaraju, C., & Prasad, M. K. (2016). Light-Weighted DNA-Based Cryptographic Mechanism Against Chosen Cipher Text Attacks. InAdvanced Computing and Systems for Security (pp. 123–144). Springer India. Babu, E. S., Nagaraju, C., & Prasad, M. K. (n. d.). Light-Weighted DNA Based Hybrid Cryptographic Mechanism Against Chosen Cipher Text Attacks. International Journal of Information Processing.

94

IPHDBCM

Bindra, G. S., Kapoor, A., Narang, A., & Agrawal, A. (2012, September). Detection and removal of cooperative blackhole and grayhole attacks in MANETs. Proceedings of the 2012 International Conference on System Engineering and Technology (ICSET) (pp. 1-5). IEEE. Capkun, S., Buttya, L., & Hubaux, J. P. (2003). Self-organized public-key management for mobile ad hoc networks. IEEE Transactions on Mobile Computing, 2(1), 52–64. Dasgupta, M., Santra, D., & Choudhury, S. (2012, November). Network Modelling of a Blackhole prevention mechanism in Mobile Ad-hoc Network. Proceedings of the 2012 Fourth International Conference on Computational Intelligence and Communication Networks (CICN) (pp. 734-738). IEEE. 10.1109/ CICN.2012.153 Deng, H., Li, W., & Agrawal, D. P. (2002). Routing security in wireless ad hoc networks. Communications Magazine, 40(10), 70–75. doi:10.1109/MCOM.2002.1039859 Gehani, A., LaBean, T., & Reif, J. (2003). DNA-based cryptography. In Aspects of Molecular Computing (pp. 167–188). Springer Berlin Heidelberg. doi:10.1007/978-3-540-24635-0_12 Gopi, A. P., Babu, E. S., Raju, C. N., & Kumar, S. A. (2015). Designing an Adversarial Model Against Reactive and Proactive Routing Protocols in MANETS: A Comparative Performance Study. Iranian Journal of Electrical and Computer Engineering, 5(5). Jain, S., Jain, M., & Kandwal, H. (2010). Advanced algorithm for detection and prevention of cooperative Black and Gray hole attacks in mobile ad hoc networks. International Journal of Computers and Applications, 1(7), 37–42. doi:10.5120/165-290 Kim, C., Talipov, E., & Ahn, B. (2006). A reverse AODV routing protocol in ad hoc mobile networks. In Emerging Directions in Embedded and Ubiquitous Computing (pp. 522–531). Springer Berlin Heidelberg. doi:10.1007/11807964_53 Konate, K., & Abdourahime, G. (2011, January). Attacks Analysis in mobile ad hoc networks: Modeling and Simulation. Proceedings of the 2011 Second International Conference on Intelligent Systems, Modelling and Simulation (ISMS) (pp. 367-372). IEEE. Kozma, W., & Lazos, L. (2009, March). REAct: resource-efficient accountability for nodemisbehavior in ad hoc networks based on random audits. Proceedings of the second ACM conference on Wireless network security (pp. 103-110). ACM. 10.1145/1514274.1514290 Kumar, S. A., Babu, E. S., Nagaraju, C., & Gopi, A. P. (2015). An Empirical Critique of On-Demand Routing Protocols against Rushing Attack in MANET. Iranian Journal of Electrical and Computer Engineering, 5(5). Leier, A., Richter, C., Banzhaf, W., & Rauhe, H. (2000). Cryptography with DNA binary strands. Bio Systems, 57(1), 13–22. doi:10.1016/S0303-2647(00)00083-6 PMID:10963862

95

IPHDBCM

Lu, S., Li, L., Lam, K. Y., & Jia, L. (2009, December). SAODV: a MANET routing protocol that can withstand black hole attack. Proceedings of the 2009 International Conference on Computational Intelligence and Security (pp. 421-425). IEEE. 10.1109/CIS.2009.244 Marti, S., Giuli, T. J., Lai, K., & Baker, M. (2000, August). Mitigating routing misbehavior in mobile ad hoc networks. Proceedings of the 6th annual international conference on Mobile computing and networking (pp. 255-265). ACM. 10.1145/345910.345955 Ning, K. (2009). A pseudo DNA cryptography method. arXiv preprint arXiv:0903.2693. Osathanunkul, K., & Zhang, N. (2011, April). A countermeasure to black hole attacks in mobile ad hoc networks. Proceedings of the 2011 IEEE International Conference on Networking, Sensing and Control (ICNSC) (pp. 508-513). IEEE. 10.1109/ICNSC.2011.5874910 Perkins, C. E., & Royer, E. M. (2001, March). The ad hoc on-demand distance-vector protocol. In Ad hoc networking (pp. 173–219). Addison-Wesley Longman Publishing Co., Inc. Raj, P. N., & Swadas, P. B. (2009). Dpraodv: A dyanamic learning system against blackhole attack in aodv based manet. arXiv preprint arXiv:0909.2371. Ramaswamy, S., Fu, H., Sreekantaradhya, M., Dixon, J., & Nygard, K. E. (2003, June). Prevention of Cooperative Black Hole Attack in Wireless Ad Hoc Networks. Proceedings of the International conference on wireless networks (Vol. 2003). Sanzgiri, K., LaFlamme, D., Dahill, B., Levine, B. N., Shields, C., & Belding-Royer, E. M. (2005). Authenticated routing for ad hoc networks. IEEE Journal on Selected Areas in Communications, 23(3), 598–610. Su, M. Y., Chiang, K. L., & Liao, W. C. (2010, September). Mitigation of black-hole nodes in mobile ad hoc networks. Proceedings of the 2010 International Symposium on Parallel and Distributed Processing with Applications (ISPA) (pp. 162-167). IEEE. 10.1109/ISPA.2010.74 Swarna, M., Umar, S., & Babu, E. S. (2016). A Proposal for Packet Drop Attacks in MANETS. In Microelectronics, Electromagnetics and Telecommunications (pp. 377-386). Springer India. doi:10.1007/97881-322-2728-1_33 Tissieres, C., Buchegger, S., & Le Boudec, J. Y. (2003). A Test-Bed for Misbehavior Detection in Mobile Ad-hoc Networks---How Much Can Watchdogs Really Do? (No. LCA-REPORT-2003-011). Tsou, P. C., Chang, J. M., Lin, Y. H., Chao, H. C., & Chen, J. L. (2011, February). Developing a BDSR scheme to avoid black hole attack based on proactive and reactive architecture in MANETs. Proceedings of the 2011 13th International Conference on Advanced Communication Technology (ICACT) (pp. 755-760). IEEE.

96

IPHDBCM

Weerasinghe, H., & Fu, H. (2007, December). Preventing cooperative black hole attacks in mobile ad hoc networks: Simulation implementation and evaluation. In Future generation communication and networking (fgcn 2007) (Vol. 2, pp. 362–367). IEEE. doi:10.1109/FGCN.2007.184 Woungang, I., Dhurandher, S. K., Peddi, R. D., & Obaidat, M. S. (2012, May). Detecting blackhole attacks on DSR-based mobile ad hoc networks. Proceedings of the 2012 International Conference on Computer, Information and Telecommunication Systems (CITS) (pp. 1-5). IEEE.

This research was previously published in the International Journal of Information Security and Privacy (IJISP), 10(3); edited by Michele Tomaiuolo and Monica Mordonini; pages 42-66, copyright year 2016 by IGI Publishing (an imprint of IGI Global).

97

Section 2

Cryptographic Algorithms

99

Chapter 6

Cryptographic Techniques Based on Bio-Inspired Systems Petre Anghelescu University of Pitesti, Romania

ABSTRACT In this chapter, bio-inspired techniques based on the cellular automata (CAs) and programmable cellular automata (PCAs) theory are used to develop information security systems. The proposed cryptosystem is composed from a combination of a CA as a pseudorandom number generator (PRNG) and a PCA that construct the ciphering functions of the designed enciphering scheme. It is presented how simple elements named „cells” interact between each other using certain rules and topologies to form a larger system that can be used to encrypt/decrypt data sent over network communication systems. The proposed security system was implemented in hardware in FPGA devices of type Spartan 3E – XC3S500E and was analyzed and verified, including NIST statistical tests, to assure that the system has good security and high speed. The experimental results proves that the cryptographic techniques based on bio-inspired algorithms provides an alternative to the conventional techniques (computational methods).

INTRODUCTION Because the communications and computer systems become each time more pervasive, cryptographic techniques plays an essential role, requiring new solutions, in order to provide data authentication, integrity and confidentiality in insecure environments. The interconnection of these pervasive devices leads to Mark Weiser’s famous vision of ubiquitous computing (Weiser, 1999) and in this way within any minute, a huge amount of information is exchanged through the Internet or over other insecure communication channels. Many kinds of information exchanges, for example text, audio/video content, in multimedia communications, should be protected from unauthorized copying, intercepting and tampering as they are traversing on public digital networks. Accordingly, cryptography has become more important in data security. Also, in the recent years, researchers have remarked the similarities between bio-inspired systems (particularly cellular automata), chaos and cryptography (Dachselt, & Schwarz, DOI: 10.4018/978-1-7998-1763-5.ch006

Copyright © 2020, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Cryptographic Techniques Based on Bio-Inspired Systems

2001; Fuster-Sabater, & Cabalerro-Gil, 2010; Kocarev, & Lian, 2011). Some of the cellular automata features as ergodicity and sensibility to the initial conditions and control parameters can be correlated with the cryptographic properties as confusion and diffusion. The essence of the theoretical and practical efforts which are done in this new field is represented by the idea that bio-inspired based encryption techniques are capable to have similar performances regarding the classic methods based on computational techniques. In this paper is presented an encryption system that uses a combination of two cellular automata: a first class of cellular automaton that generates the evolution rules for the second class of five programmable cellular automata arranged in pipeline. The entire security system was implemented both in software using C# programming language and in hardware on a FPGA of type Spartan 3E – XC3S500E in which the plaintext/ciphertext is received/transmitted using User Datagram Protocol (UDP). This chapter is organized in eight sections. In the background section, are described some basic theoretical foundations of the proposed work that includes CAs and PCAs. The third section, provides a brief overview of the classical cryptography and bio-inspired systems in cryptography. The next section, on reconfigurable hardware devices, introduces the existing reconfigurable hardware devices approaches for supporting bio-inspired algorithms and presented also the reasons for using them in the application presented in this chapter. Then, the section bio-inspired based algorithm for cryptography, describes the proposed bio-inspired encryption algorithm used to encrypt and decrypt data sent over the communication networks. Additionally, in section testing and experimental results, are made the investigations of statistical properties of the encrypted sequences (performed using NIST statistical tests), distribution of text (plaintext and ciphertext) and encryption/decryption speed. In the next section, are presented the future research direction of the research presented in this chapter. Finally, section eight, conclude the chapter.

BACKGROUND The intersection of biology and computer science has been a productive field for some time. On one hand, CAs is a bio-inspired paradigm highly addressing the soft computing and hardware for a large class of applications including information security. On the other hand, PCAs is a modified CAs structure including switches in order to allow the self-organizing of the cellular structure.

Cellular Automata (CA) CAs, first introduced by von Neumann and Stanislav Ulam (Neumann, 1966) in the ‘50s, exhibit useful and interesting characteristics and has attracted researchers from different field of interests, who applied it in different ways. The most notable characteristics of CAs are: massive parallelism, locality of cellular interactions and simplicity of basics components. CAs perform computations in a distributed way on a spatial grid and differ from a standard approach to parallel computations whereby a problem is split into independent sub-problems later to be combined in order to yield a final solution. CAs suggest a new approach in which a complex global behavior can be modelled by non-linear spatially extended local interactions. Thus far, CAs have been used primarily to model the systems consisting of a number of elements obeying identical laws of local interactions (e.g. problems of fluid dynamics, plasma physics, chemical systems, crystals growth, economics, two-directional traffic flow, image processing and pattern recognition, 100

Cryptographic Techniques Based on Bio-Inspired Systems

parallel processing, random number, evolution of spiral galaxies, modeling of very complicated physical or chemical processes, molecular computing) (Adamatzky, 1994; Ilachinski, 2001; Wolfram, 2002). The wide applicability of CA is somehow limited because the methodologies for designing CA, intended to solve specific predefined tasks, are still underdeveloped. Such designing techniques would be extremely useful since there exist many problems for which local interactions, that would drive CA to solve these problems, are not known in advance. To the best of my knowledge, some works has already been done in this area, mainly using genetic algorithms to find evolution rules for the specific CAs. CAs are mathematical idealizations of physical systems in terms of discrete time and space, where interactions are only local and where each cell can assume the value either 0 or 1 (Sung-Jin, 2004). In fact, CA represents a particular class of dynamical systems that enable to describe the evolution of complex systems with simple rules, without using partial differential equations. Each cell of the CA is restricted to local neighborhood interactions only, and as a result it is incapable of immediate global communication. The neighborhood of the cell is taken to be the cell itself and some or all of the immediately adjacent cells. The CA evolves in discrete steps, with the next value of one site determined by its previous value and that of a set of sites called the neighbor sites. The extent of the neighborhood can vary, depending among other factors upon the dimensionality of the CA under consideration. Classical examples for cell neighborhoods are presented in Figure 1 (left side, von Neumann Neighborhood - with 3 cells for one-dimensional CA respective 5 cells for bi-dimensional CA and right side, Moore neighborhood with 3 cells for one-dimensional cellular automata respective 9 cells for bi-dimensional cellular automata considers both kinds the direct and the diagonal neighbors). The state of each cell is updated simultaneously at discrete time steps, based on the states in its neighborhood at the preceding time step. The algorithm used to compute the next cell state is referred to as the CA local rule. Typically, a CA consists of a graph where each node is a finite state cell. This graph is usually in the form of a two-dimensional lattice whose cells evolve according to a global update function applied uniformly over all the cells. As arguments, this update function takes the cell’s present state and the states of the cells in its interaction neighborhood as shown in Figure 2. Figure 1. CA classical neighborhoods (a) von Neumann neighborhood, (b) Moore neighborhood

101

Cryptographic Techniques Based on Bio-Inspired Systems

Figure 2. CAs state transitions depending on the neighborhood states

The next-state function describing a rule for a three neighborhood CA cell where assuming that i is position of an individual cell in an one dimensional array, t is time step, and ai(t): output state of the i-th cell at the t-th time step can be expressed as follows: ai (t + 1) = f ai (t ), ai +1(t ), ai −1(t )

(1)

where f denotes the local transition function realized with a combination logic and is known as a rule of the CA. If the rule of a CA involves only XOR logic, then it is called a linear rule. Rules involving XNOR logic are referred to as complement rules. A CA with all its cells having linear rules is called a linear CA, whereas a CA having a combination of XOR and XNOR rules is called additive CA. If all the cells obey the same rule, then the CA is said to be a uniform CA, otherwise, it is a hybrid CA. A CA is said to be a null boundary CA if both the left and right neighbor of the leftmost and rightmost terminal cell is connected to logic 0-state. For example, in case of one-dimensional, three neighborhood and two-state cell, the number of all possible uniform CA rules is 256 (28). These rules are enumerated using Wolfram’s naming convention (Wolfram, 1986) from rule number 0 to rule number 255 and can be represented by a 3-variable Boolean function. As an example, in Table 1, are presented five fundamental rules that are obtained using 3-neighborhood (the cell on the left, the cell itself and the cell from the right side of the cell in question). This means that the rule for this three cell neighborhood must contain 8 bits. These rules, arranged in a certain mode, are used in the proposed encryption system presented in later sections to construct the cryptographic algorithm. The space of evolution rules depends on the number of possible states of the current cell and the number of its neighbors. This leads to an exponential growing of the rules space (see Table 2).

102

Cryptographic Techniques Based on Bio-Inspired Systems

Table 1. An example of CA evolution rules construction Rules (Decimal Number)

7 111

6 110

5 101

4 100

3 011

2 010

1 001

0 000

90

0

1

0

1

1

0

1

0

150

1

0

0

1

0

1

1

0

51

0

0

1

1

0

0

1

1

60

0

0

1

1

1

1

0

0

102

0

1

1

0

0

1

1

0

2

2

2

2

2

2

2

20

7

6

5

4

3

2

1

Table 2. The size of the rules in function of neighborhood dimension Size of the Cell Neighborhood

Decimal Rule Size (Maximum Value)

Rule Bit Size

2

16

4

3

256

8

4

65536

16

5

4294967296

32

6

18446744073709551616

64

7

3.40E+38

128

8

1.16E+77

256

9

1.34E+154

512

The systematic study of CA was initialized by S. Wolfram in (Wolfram, 2002) in which he studied the relationships between CA and different dynamical systems. According to (Wolfram, 2002) there are four classes of CA whose comportment can be compared with the similar behavior of the dynamic systems (given in parenthesis): Class I: The CA evolution from all initial configurations reaches the same final state and stays there (limit points). Class II: The CA encounters simple or cyclic structures, but which one depends on the initial configuration (limit cycles). Class III: The CA from majority of initial states lead to arbitrary patterns (chaotic behavior of the kind associated with strange attractors). Class IV: The CA from some of initial configurations generates global complex structures (very long transients with no apparent analog in continuous dynamic systems). This behavior basically means that CA can be shown to be capable of universal computation. This classification mainly refers to 1-D CAs, but similar ones can be made for 2-D or 3-D cases. Classifications of CA are done by means of empirical observations of CA evolutions (space-time patterns). The very large phenomenology of the CA model, its apparently big complexity and massive parallelism (however, this parallelism, when emulated in software or in sequential hardware, disappears) offer a good basis for applications in cryptography and not only. 103

Cryptographic Techniques Based on Bio-Inspired Systems

Programmable Cellular Automata (PCA) The programmable cellular automata (PCA) was firstly introduced in (Nandi et al., 1994) as a modified CA structure. These PCA allows spatial and temporal variations in the state transition rules within a CA according to some external control scheme and signals. Practically, PCA dynamically change the CA rules. As an example, using such a cell structure as in Figure 3, all possible non-complemented additive rules can be achieved through the combinations of the control signals of C1, C2 and C3 switches. Thus the PCA architecture is very much flexible and enable to emulate via control signals different hybrid CA configurations (Anghelescu et al., 2010). In practice, a control program, stored in memory, can be employed to activate the switches. For example, the state 1 or 0 of the bit i-th of a memory word either opens or closes the switch that controls that cell. Basically, such a structure is referred as a PCA. In this paper, PCA is used to implement the proposed encryption algorithm. All in all, the CAs and PCAs offer a good basis for applications in cryptography and represent a “converse pole” as computing architecture in comparison with sequential model: are parallel systems without central processing unit in which the computation powerful of his elements are much reduced.

LITERATURE REVIEW In present, promising applications for cryptographic algorithms may be classified into two categories: Category I: Processing of large amount of data at real time potentially in a high speed network. Examples include telephone conversations, telemetry data, video conferencing, streaming audio or encoded video transmissions and so forth. Figure 3. PCA with 3-neighbors and all non-completed additive evolution rules

104

Cryptographic Techniques Based on Bio-Inspired Systems

Category II: Processing of very small amount of data at real time in a moderately high-speed network transmitted unpredictably. Examples include e-commerce or m-commerce transactions, bank account information extraction, e-payments and micro-browser-based (WAP-style), HTML page browsing and so forth. Cryptographic realizations could be done in software or in hardware (Henriquez et al., 2006). In software platforms can be used for those security applications where the data traffic is not too large and thus low encryption rate is acceptable. On the other hand, hardware methods offer high speed and bandwidth, providing real time encryption if needed. A good overview on the all major cryptographic techniques can be found in reference (Menezes et al., 1996; Stallings, 2003; Cusick & Stanica, 2009; Koc, 2009). Accordingly, the cryptographic methods are divided into two categories: symmetric-key (or secret-key) and asymmetric-key (or public-key). In symmetric cryptography the same key is used for both encryption and decryption, whereas in asymmetric cryptography there are two keys: one for encryption (which is public known), and other for decryption (which must remain secret). Based on these algorithms there are mainly two classes of symmetric-key encryption schemes: block ciphers and stream ciphers. Block ciphers breaks up the message into blocks of the fixed length and encrypt one block at a time. On the contrary, the stream ciphers encrypt a single bit of plain text at a time. Encryption is accomplished by combining the cipher sequence with the plaintext, usually with the bitwise XOR logic operation. There are several methods in order to generate the cipher sequence beginning from the secret key: LFSR (Linear Feedback Shift Register) generator, BBS (Blum Blum Shub) generator, and so forth (Menezes et al., 1996). As we said above, there are several mathematical techniques that can be used for cryptographic purpose, and one of them is the use of bio-inspired systems or discrete dynamical systems (Schmitz, 2001). In the domain of researches having as subject the association between the cellular automata and cryptography was reported more encryption systems based on the cellular automata theory. CA were proposed for both secret-key and public-key cryptography. Remarkable is the fact that the relationship between the CA and cryptography was revealed by Shannon in his fundamental early work (Shannon, 1949): Good mixing transformations are often formed by repeated products of two simple non-commuting operations. Hopf has shown, for example, that pastry dough can be mixed by such a sequence of operations. The dough is first rolled out into a thin slab, then folded over, then rolled, and the folded again, etc. It seems that Shannon discusses about a system composed from simple components that interaction between them – with a transparent local comportment – but the global comportment of the entire system unsuspected, things that are well known in the CA theory. So that, a very simple variant used for encryption using cellular automata is reported by Stephen Wolfram in (Wolfram, 1986) and (Wolfram, 2002) and is based on the fact that the CA from class III (conform the Wolfram classification) are dynamical chaotic systems. In this case, the evolution of the cellular automaton depends considerable of the initial state, but we can say that after some time the state is forgotten in sense in which cannot be found from current configuration analyses. Anyway, if we repeat the initial state, the evolution will be the same. Wolfram use a uniform 1D CA with three neighborhoods, and rule 30 to generate pseudorandom number sequences (CA-PRNG). The encryption system proposed

105

Cryptographic Techniques Based on Bio-Inspired Systems

by Wolfram can be included in category “Chaotic stream ciphers based on the pseudorandom number generator” (PRNG). The based principle of these ciphers is to obtain the encryption text by mixing the output of these pseudorandom number generators with the message (Lee et al., 2003; Koc, 2009). Another solution used for encryption with cellular automata is presented in (Hortensius et al., 1989) and (Nandi et al., 1994). These used non-uniform CA with two rules 90 and 150, and it was found that the quality of pseudorandom number generated was better that the quality of the Wolfram system. (Tomassini & Perrenoud, 2000) proposed to use non-uniform, 1D CAs with four rules 90, 105, 150 and 165, which provide high quality pseudorandom number sequences and a huge space of possible secret keys which is difficult for cryptanalysis. Another variant of encryption system based on the cellular automata, which consider also the inverse iteration, is presented in (Adamatzky, 1994) and (Martin, 2004). Here is used a bi-dimensional cellular automaton and the dates are the initial state of the cellular automaton. Using a reversible evolution rule the initial message is modified progressive. The message is decrypted rolling the inverse rule the same number of iterations as to encryption. This encryption system can be included in category of “Stream ciphers based on inverse iteration-with reaction”. These systems can be also based on a series of evolution rules that served as chaotic system, rules used for encryption and decryption (Lu et al., 2004). Other cryptosystem realized with the help of cellular automata combine the direct and inverse iteration (Gutowitz, 1994), “Block ciphers based on the direct and inverse iteration” (Masuda & Aihara, 2002) – these ciphers was as a general rule proposed for image encryption. Here is used a bi-dimensional cellular automaton, the message being the initial state of this. The codification implied the inverse iteration of a rule, the key is a rule. This rule is not necessary to be reversible: for inverse iteration is chosen randomly one of the possible states of the cellular automaton. For decryption we must know the rule that is direct iterated and use the same number of steps (Anghelescu et al., 2010; Anghelescu et al., 2013). (Tripathy & Nandi, 2009) have designed a lightweight CA-based symmetric-key encryption that supports 128-bit block size with 128-, 192- and 256-bit keys. The motivation for embarking on CA for their cipher design is due to the fact that CA provides a high level of parallelism and therefore, able to achieve high speeds. The cipher has also been proved to be resistant against timing analysis attacks. A hybrid CA with 2 rules 30 and 134 is proposed as cryptographic hash functions in (Jamil et al., 2012), in which elementary CA rules are used for mixing bits of the message. The cipher has also been analyzed using NIST tests and passes all the statistical tests. The deterministic chaotically dynamic of CA is similar with pseudorandom systems used in classical methods and the complexity of a dynamical system that determine the efficiency of the entire information protection process is equivalent with the algorithm complexity from classical encryption methods. Wolfram (Wolfram, 2002) pointed out that the future researches in this domain must be oriented to find solutions to complete the classical encryption systems with the other based on the bio-inspired systems (for example cellular automata theory), and less to design work of new methods developed “ad-hoc”. From the above discussion, seeing the promised potential of using of the CA in the field of cryptography, we believe that research on bio-inspired cryptography will be helpful to benefit the conventional cryptology and open a broader road to the design of ciphers with very good properties.

106

Cryptographic Techniques Based on Bio-Inspired Systems

RECONFIGURABLE HARDWARE AND CA The reconfigurable devices as piece of hardware able to dynamically adapt to algorithms and become through electrically program, almost any kind of digital circuit or system, was firstly introduced by G. Estrin in 1960. His invention consists on a hybrid machine composed by a general purpose microprocessor interconnected with programmable logic devices (Bobda, 2007). In general, reconfigurable devices tend to be a good choice when dealing with algorithms that implies high parallelism as CA based algorithms are. VLSI (also known as ASIC - Application Specific Integrated Circuit) and FPGAs (Field Programmable Gate Arrays) are two distinct alternatives for implementing cryptographic algorithms in hardware. An electronic device is said to be configurable (or programmable) when its functionality is not pre-defined at fabrication-time, but can be further specified by a configuration bit-stream (or a program). Reconfigurable devices permit configuration several times, supporting system upgrades and refinements in a relatively small time scale. Given this architectural flexibility and upgradeability, they constitute the best candidate for supporting bio-inspired architectures: they offer a set of features that permit the implementation of flexible architectures, while still guaranteeing high performance execution (Upegui, 2010). A FPGA consists of an array of logic elements together with an interconnect network which can be configured by the user at the point of application. The basic structure of a FPGA circuit that contain an array/matrix of configurable logic blocks of potentially different types, including general logic, memory and multiplier blocks, surrounded and connected via programmable interconnects is presented in Figure 4. User programming specifies both the logic functions of each block and the connections between the blocks. In one sense, FPGAs represent an evolutionary improvement in gate array technology which offers potential reductions in prototype system costs and product time-to-market. However, recent applications of FPGA technology suggest their impact on electronic systems may be much more profound. Consider the fact that reprogrammable FPGAs are capable of dynamically changing their logic and interconnect structure to adapt to changing system requirements. This offers a new computing paradigm which blurs the traditional lines between hardware and software. The main advantage of FPGAs is their re-configurability, i.e., they can be used for different purposes at different stages of a computation and they can be, at least partially, reprogrammed on run-time. The two most popular FPGA manufacturers are Xilinx and Altera. Today, due to cost decreases and the flexibility of a programmable solution, highdensity FPGAs are often employed for system-level prototyping. All in all, the using of FPGA devices for bio-inspired based cryptosystems is motivated by four reasons: encryption speed (encryption algorithms based on multiple cells interconnected contain many operations on that uses communications between adjacent cells and these can be executed in parallel or pipelined mode in hardware, while not suitable in software); security assured (there is no physical protection for an encryption algorithm written by software and an intruder can go in with various debugging tools to modify the algorithm without anyone ever realizing it; hardware encryption devices can be securely encapsulated to prevent this.); flexibility (all the implemented functions could be upgraded at any time); analogy between the re-configurability and CA (all the CA cells could be updated in parallel, in a single time clock).

107

Cryptographic Techniques Based on Bio-Inspired Systems

Figure 4. Basic FPGA structure

BIO-INSPIRED BASED ALGORITHM FOR CRYPTOGRAPHY In this section is presented the proposed cryptosystem used to encrypt/decrypt data sent over the communication networks. This cryptosystem uses a one-dimensional CA and a combination of five onedimensional PCAs arranged in pipeline and a control logic that manage all the operations of the CA and PCAs. The basic structure of the cryptosystem is presented in Figure 5. Because a lot of simulations and research has been carried out using 8-bit PCAs, an 8-bit PCAs was chosen for this project. In the block cipher scheme, one 8-bit message block is enciphered by one enciphering function. The enciphering function has five fundamental transformations (FTs) in order to operate on 8-bit data. This FTs are constructed using five PCAs arranged in pipeline. The block cipher scheme can be mathematically expressed as follows: If M is a block of text (8 bits plaintext), C is a block of text (8 bits ciphertext), E is an enciphering function

108

Cryptographic Techniques Based on Bio-Inspired Systems

Figure 5. Basic structure of the cryptosystem

Then encryption: C = E •M

(2)

Decryption: M = E −1 • C

(3)

where: ● denotes a function symbol. The algorithm discussed here to encrypt/decrypt data sent over communication networks can be divided into two-phases: an encryption phase and a decryption phase. In the encryption phase the initial PCA configurations that practically contain inside the plaintext are evolved a number of predefined steps (between 1 and 7). The PCAs control signals are generated using the interconnections with the first CA that acts like a pseudo-random number generator (PRNG). The encryption phase for the pipelined PCAs block cipher is presented in Figure 6. In the decryption phase the initial PCA configurations that contain inside the ciphertext are evolved a number of steps that must be accordingly with the differences between 8 and number of steps used for encryption phase. The diagram of the decryption phase for the pipelined PCAs block cipher is presented in Figure 7. In the decryption phase decrEvolSteps is 8 - encrEvolSteps because the PCAs discovered and used in this cryptosystem generates for any initial state (that could be plaintext or ciphertext) cycles of even length that will repeat themselves after 8 steps. The PCAs used in this cryptographic algorithm is a hybrid PCA and is configured with the combination of the rules 51, 60 and 102 presented in the Table 1. In Figure 8 is presented an example of cycles generated by the PCA configured with rules cell 1 – rule 60, cell 2 – rule 51, cell 3 – rule 60, cell 4 – rule 60, cell 5 – rule 60, cell 6 – rule 51, cell 7 – rule 51, cell 8 – rule 51 and initial states 128 in the left side and 242 in the right side. As it is shown in Figure 8, the PCA has two equal cycles of length 8. This property is a basic requirement of the cryptographic scheme. For example, in the enciphering phase, if it is used this PCA as enciphering function and define a plaintext as its initial state, it goes to its intermediate state after four cycles. In the deciphering phase, after running another four cycles, the intermediate state returns back to its initial state, so the cipher text is deciphered into plaintext. Because the PCA does not generate se-

109

Cryptographic Techniques Based on Bio-Inspired Systems

Figure 6. The diagram of the encryption phase

110

Cryptographic Techniques Based on Bio-Inspired Systems

Figure 7. The diagram of the decryption phase

111

Cryptographic Techniques Based on Bio-Inspired Systems

Figure 8. The transitions diagram of the PCA with initial states 128 (left side) and 242 (right side)

quences of maximum-length for all the possible combinations (512) of the rules it is necessary to apply from the first CA used as PRNG only the combinations of the three rules 51, 60 and 102 that generate cycles of length 8. The rules with 8-cycle length are only 156 and are presented in detail in my previous paper (Anghelescu et al., 2010). In concordance with the CA theory, a single PCA cell was designed and then was multiplied in the FPGA circuits. The cell consists of a D flip-flop and a logic combinational circuit (LCC). The LCC includes multiplexers and XNOR logic gates to implement the rules of CA and to control the loading of data and operation of the CA. The entire scheme of that cell is presented in detail in paper (Anghelescu et al., 2013). The CA that select which rules are applied to the cells of the PCAs is realized by using a combination of rules 90 and 150 presented in Table 1. It has established in paper (Hortensius & Podaima, 1990) that the maximum-length CAs with rules 90 and 150 generates patterns having a high quality of pseudo-randomness. The control logic is the heart of the design and includes the communication interface between the FPGA and PC using UDP protocol and controls all operations of the cryptosystem. UDP is a simple to implement protocol because it does not require keeping track of every packet sent or received and it does not need to initiate or end a transmission.

TESTING AND EXPERIMENTAL RESULTS A lot of experiments were carried out throughout the development of the cryptosystem, motivated by three reasons: First: For investigation of the CA and PCA comportment, Second: For assuring that the generated sequences respect the CA and PCA theory and Third: To assure that the pipeline PCA based cryptosystem has a good security and high speed. The general structure of the cryptosystem is presented in Figure 9. In order to put in practice the principle of working of the cryptosystem was proposed an architecture which is able to satisfy both of the requirements for communication and for specifications related on PCA. FPGA circuits are used to implement the pipeline PCA block cryptosystem. The cryptosystem is implemented using VHDL code and is written in Xilinx ISE and Active HDL using structural and behavioral specifications. The hardware implementation of the PCA cryptosystem was realized using a Spartan 3E XC3S500E FPGA board from Xilinx (Xilinx, 2011) (see Figure 10).

112

Cryptographic Techniques Based on Bio-Inspired Systems

Figure 9. Bio-inspired cryptosystem: general architecture

Figure 10. FPGA board used for test the bio-inspired based cryptosystem

The FPGA board is interfaced with a host computer using RJ-45 connector and UDP protocol. The UDP allows high speed data transfer from the PC to the cryptosystem and inverse. In Figure 11 is presented an illustrative example for the cryptographic phases (encryption and decryption process) applied to a text file. The message (plaintext or ciphertext) is divided into 1Kb packages and is sent to the FPGA board using the UDP client – server connection. As the bytes reaches destination they are immediately encrypted using the correspondent bytes of the PCA’s state and then saved into the 1Kb RAM memory of the board. In the FPGA, the message received is treated character by character as we explained above and the encryption/decryption results are sent back by the FPGA to the PC to be displayed, stored and analyzed. In hardware, the encryption rules are applied using the CA as PRNG or are downloaded to the RAM memory before encryption. When the encryption process begins, rules are generated or read out in sequence and sent to the PCAs. The process of generated or read rules does not introduce delays in the cryptographic algorithm because are generated/read in parallel with the encryption/decryption of a block of message.

113

Cryptographic Techniques Based on Bio-Inspired Systems

Figure 11. Results for encryption and decryption phases

To show that the bio-inspired based cryptosystem has the property of confusion specific for classical cryptography it is possible to identify the distribution of the plaintext and ciphertext in ASCII intervals. This cryptosystem will map the given plaintext into a random ciphertext, which means that no pattern appears in the ciphertext. In the Figure 12 and Figure 13, the distribution of the ASCII values for a plaintext and ciphertext is presented. In Figure 12, plaintext distribution, most of the characters are lowercase, so the distribution is dense in the interval 97 and 122. In Figure 13, ciphertext distribution, the ciphertext is distributed almost uniform in the complete interval of ASCII values, and not only in the zone of alphanumeric intervals, and hence the developed cryptosystem also maps plaintext to a random ciphertext. In the PCA encryption algorithm, the same ciphertext may be generated from different plaintext, and any ciphertext may give rise as well to different plaintext depending on the different PCA’s rule configuration. The PCA based encrypted sequences was tested using a set of 16 statistical tests conceived by the National Institute of Standards and technology (NIST) (Rukhin et al., 2010). The NIST test generates probabilistic results with respect to some characteristics that describe the pseudo-random number generators. The encrypted sequences pass the NIST tests and the system is accepted as possible random (Table 3).

114

Cryptographic Techniques Based on Bio-Inspired Systems

Figure 12. Distribution of the plaintext in ASCII interval

Figure 13. Distribution of the ciphertext in ASCII interval

115

Cryptographic Techniques Based on Bio-Inspired Systems

Table 3. Results for NIST statistical tests Crt. No.

Test Applied

Crt. No.

Interpretation

Test Applied

Interpretation

1.

Frequency Test

PASS

9.

Maurer’s Universal Statistical

PASS

2.

Frequency Test within a Block

PASS

10.

Linear Complexity Test

PASS

3.

Runs Test

PASS

11.

Serial Test

PASS

4.

Test for the Longest Run of Ones in a Block

PASS

12.

Approximate Entropy Test

PASS

5.

Binary matrix Rank Test

PASS

13.

Cumulative Sums Forward

PASS

6.

Discrete Fourier Transform

PASS

14.

Cumulative Sums Reverse

PASS

7.

Non-Overlapping Template Matching Test

PASS

15.

Random Excursions Test

PASS

8.

Overlapping Template Matching Test

PASS

16.

Random Excursions Variant

PASS

The encrypted sequences pass the NIST tests and the cryptosystem is accepted as possible random. The timing analyzer was used to determine the critical path and the maximum operating frequency (33Mbps at 50MHz CLK Xilinx FPGA – XC3S500E). The encryption system discussed above can be applied in data communication networks, both in private network (Local Area Network – LAN) and public network (Wide Area Network – WAN).

FUTURE RESEARCH DIRECTIONS In the immediate future, in order to assure flow control and guarantees that all sent packets will reach the destination in the correct order, the bio-inspired based cryptosystem will be implemented using the TCP protocol instead of UDP protocol. Also, the FPGA board will be connected to a router in order to assure encryption and decryption of messages from any computer connected to the Internet. The value for encryption/decryption speed will be improved by using increased FPGA clock speed and increased space RAM memories in order to store more encrypted/decrypted TCP/UDP packages into the FPGA before starting sending back the messages to the PC in the transmission phase.

CONCLUSION In the present work, is demonstrated the ability of bio-inspired based algorithms to combine the two necessary properties diffusion (sensitivity to the initial conditions and/or control parameters) and confusion (the output has the same distribution for any plaintext used like input) in order to obtain cryptographic solutions. The major aspect treated in this paper is the cryptosystem description based on the CA and PCA paradigms, where an original model is proposed to cooperate and create a cryptographic system. The basic algorithms for each encryption and decryption involved in the cryptosystem architecture are described, too. According to the results obtained with respect to security analysis, the proposed method is expected to be useful for real-time encryption/decryption and transmission applications. The general

116

Cryptographic Techniques Based on Bio-Inspired Systems

conclusion of this work is that it is possible to build evolutionary cryptosystems based on a simple mathematical models specific of CA and PCA by introducing the local interaction between cells, local evolution rules, and massive parallelism.

REFERENCES Adamatzky, A. (1994). Identification of Cellular Automata. London: Taylor & Francis Ltd. Anghelescu, P., Ionita, S., & Iana, V. (2013). High-speed PCA Encryption Algorithm using Reconfigurable Computing. Journal of Cybernetics and Systems, 44(4), 285–304. Anghelescu, P., Ionita, S., & Sofron, E. (2010). Encryption Technique with Programmable Cellular Automata. Journal of Cellular Automata, 5(1-2), 79–106. Bobda, C. (2007). Introduction to Reconfigurable Computing – Architectures, algorithms and applications. Springer. Cho, S.-J., Choi, U.-S., Hwang, Y.-H., Kim, H.-D., Pyo, Y.-S., Kim, K.-S., & Heo, S.-H. (2004). Computing Phase Shifts of Maximum-Length 90/150 Cellular Automata Sequences. Proceeding of the 6th International Conference on Cellular Automata for Research and Industry, ACRI 2004, Amsterdam (pp. 31-39). 10.1007/978-3-540-30479-1_4 Cusick, T., & Stanica, P. (2009). Cryptographic Boolean functions and applications. Elsevier. Dachselt, F., & Schwarz, W. (2001). Chaos and Cryptography. IEEE Transactions on Circuits and Systems, 48(12), 1498–1509. doi:10.1109/TCSI.2001.972857 del Rey, Á.M. (2004). A Novel Cryptosystem for Binary Images. Grant SA052/03, Studies in Informatics and Control, 13(1). Fuster-Sabater, A., & Cabalerro-Gil, P. (2010). Chaotic Cellular Automata with Cryptographic Application. Proceedings of the 9th International Conference on Cellular Automata for Research and Industry (pp. 251–260). Springer-Verlag Berlin Heidelberg Gutowitz, H. (1994). Methods and Apparatus for Encryption, Decryption and Authentication using Dynamical Systems. Rodriguez-Henriquez, F., Saqib, N.A., Días Pérez, A.D., & Koc, C.K. (Eds.), (2006). Cryptographic algorithms on reconfigurable hardware. Springer Science + Business Media, LLC. Hortensius, P. D., McLeod, R. D., & Card, H. (1989). Parallel Random Number Generation for VLSI Systems using Cellular Automata. IEEE Transactions on Computers, 38(10), 1466–1473. doi:10.1109/12.35843 Hortensius, P. D., & Podaima, R. D. (1990). Cellular Automata Circuits for Built-in Self-test. IBM Journal of Research and Development, 34(2/3), 389–405. doi:10.1147/rd.342.0389 Ilachinski, A. (2001). Cellular Automata – A Discrete Universe. Singapore: World Scientific Publishing Co. Pte. Ltd.

117

Cryptographic Techniques Based on Bio-Inspired Systems

Jamil, N., Mahmood, R., & Muhammad, R. (2012). A New Cryptographic Hash Function Based on Cellular Automata Rules 30, 134 and Omega-Flip Network. Proceedings of the 2012 International Conference on Information and Computer networks (ICICN 2012), Singapore (Vol. 27, pp. 163-169). IACSIT Press. Koc, C.K. (Ed.), (2009). Cryptographic engineering. Springer Science + Business Media, LLC. doi:10.1007/978-0-387-71817-0 Kocarev, L., & Lian, S. (2011). Chaos-Based Cryptography – Theory, Algorithms and Applications. Springer-Verlag Berlin Heidelberg. doi:10.1007/978-3-642-20542-2 Lee, P. H., Pei, S. C., & Chen, Y. Y. (2003). Generating chaotic stream ciphers using chaotic systems. The Chinese Journal of Physiology, 41, 559–581. Lu, H., Wang, S., Li, X., Tang, G., Kuang, J., Ye, W., & Hu, G. (2004). A new spatiotemporally chaotic cryptosystem and its security and performance analyses. Chaos (Woodbury, N.Y.), 14(3), 617–629. doi:10.1063/1.1772731 PMID:15446972 Masuda, N., & Aihara, K. (2002). Cryptosystems with discretized chaotic maps. IEEE Trans. Circuits Syst. I, 49(1), 28–40. doi:10.1109/81.974872 Menezes, A., Oorschot, P., & Vanstone, S. (1996). Handbook of applied cryptography. CRC Press. doi:10.1201/9781439821916 Nandi, S., Kar, B. K., & Chaudhuri, P. P. (1994). Theory and applications of cellular automata in cryptography. IEEE Transactions on Computers, 43(12), 1346–1356. doi:10.1109/12.338094 Rukhin, A., Soto, J., Nechvatal1, J., Smid, M., Barker, E., Leigh, S., Levenson, M., Vangel, M., Banks, D., Heckert, A., Dray, J., Vo, S. (2010). A Statistical Test Suite for Random and Pseudorandom Number Generators for Cryptographic Applications (NIST Special Publication 800-22). Schmitz, R. (2001). Use of Chaotic Dynamical Systems in Cryptography. Journal of the Franklin Institute, 338(4), 429-441. Shannon, C. (1949). Communication Theory of Secrecy Systems. Bell Sys. Tech. J., 28, 656–715. Retrieved from netlab.cs.ucla.edu/wiki/files/shannon1949.pdf Stallings, W. (2003). Cryptography and Network Security (3rd ed.). Prentice Hall. Tomassini, M., & Perrenoud, M. (2000). Stream Ciphers with One- and Two-Dimensional Cellular Automata. In M. Schoenauer at al. (Eds.), Parallel Problem Solving from Nature - PPSN VI, LNCS (Vol. 1917, pp. 722-731). Tripathy, S., & Nandi, S. (2009). LCASE: Lightweight cellular automata-based symmetric key encryption. International Journal of Network Security, 8(2), 243–252. Upegui, A. (2010). Dynamically Reconfigurable Hardware for Evolving Bio-Inspired Architectures. In R. Chiong (Ed.), Intelligent Systems for Automated Learning and Adaptation: Emerging Trends and Applications (pp. 1–22). doi:10.4018/978-1-60566-798-0.ch001 von Neumann, J. (1966). Theory of Self -Reproducing Automata (Ed., Burks, A.W.). London: Univ. of Illinois Press.

118

Cryptographic Techniques Based on Bio-Inspired Systems

Weiser, M. (1999). The computer for the 21st century. Mobile Computing and Communications Review, 3(3), 3–11. doi:10.1145/329124.329126 Wolfram, S. (1986). Cryptography with Cellular Automata. Springer, Advances in Cryptology: Crypto ’85 Proceedings. LNCS, 218, 429–432. Wolfram, S. (2002). A new kind of science. Champaign, IL: Wolfram Media Inc. Xilinx. (2011). Spartan 3E Starter kit board data sheet. Retrieved from http://www.xilinx.com/support/ documentation/boards_and_kits/ug230.pdf

This research was previously published in the Handbook of Research on Advanced Hybrid Intelligent Techniques and Applications edited by Paramartha Dutta, Pinaki Banerjee, Dipankar Majumdar, and Siddhartha Bhattacharyya; pages 59-77, copyright year 2016 by Information Science Reference (an imprint of IGI Global).

119

120

Chapter 7

Modification of Traditional RSA into SymmetricRSA Cryptosystems Prerna Mohit Indian Institute of Technology (Indian School of Mines), India G. P. Biswas Indian Institute of Technology (Indian School of Mines), India

ABSTRACT This paper addresses the modification of RSA cryptography namely Symmetric-RSA, which seem to be equally useful for different cryptographic applications such as encryption, digital signature, etc. In order to design Symmetric-RSA, two prime numbers are negotiated using Diffie-Hellman key exchange protocol followed by RSA algorithm. As the new scheme uses Diffie-Hellman and RSA algorithm, the security of the overall system depends on discrete logarithm as well as factorization problem and thus, its security is more than public-key RSA. Finally, some new cryptographic applications of the proposed modifications are described that certainly extend the applications of the existing RSA.

1. INTRODUCTION As the popularity of internet technology is increasing, more and more numbers of users are using it. In addition, its security protection is also very important. Hence, to provide security protection, data need to be encrypted before transmitting over a public channel. Two basic technologies are used for the encryption of data, i.e. symmetric key encryption and asymmetric key encryption (Mohit et al., 2015; Mohit et al., 2016; Sun et al., 2007). One of the very well-known cryptography scheme is RSA algorithm (Sun et al., 2007; Peng et al., 2016; Ambedkar et al., 2011; Minni et al., 2013). In broad terms the security of traditional public-key cryptosystem either depend on the factorization problem or the discrete logarithm problem. The RSA-type algorithm comes under the factorization problem and the Diffie-Hellman (DH) comes under discrete logarithm problem. We modify the RSA encryption technology using a combination DOI: 10.4018/978-1-7998-1763-5.ch007

Copyright © 2020, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Modification of Traditional RSA into Symmetric-RSA Cryptosystems

of DH, RSA and presented a more efficient encryption/decryption technology. In 1998, Takagi (Takagiet al., 1998) showed the extension of RSA to multi-prime algorithm with modulus pkq and reduced the decryption time by using Quisquater-Couvreur method. After a few years in 2002, Elkamchouchi et al. (Elkamchouchi et al., 2002) proposed extended RSA, where RSA algorithm is implemented using Gaussian integers over real and imaginary numbers. Later on, several versions of the extended RSA are developed that shows the validity of executing RSA algorithm over the complex numbers (El-Kassar et al. 2005) (Verkhovsky et al., 2011), however its competence is lesser than the existing RSA. In 2015 M. Thangavel (Thangavel, 2015) proposed an enhanced version of RSA, which uses several parameters such as four prime numbers, three Euler functions, multiple public and private exponents that increase the overall computational complexity and overhead of the scheme. In this paper, a modification of the RSA algorithm is proposed. The public-key RSA is converted into a Symmetric-key RSA (SYM-RSA) cryptosystem, where two prime numbers are securely exchanged using a hybrid of Diffe-Hellman key exchange, RSA and used them for the generation of a secret key between two participants. The rest of the paper is presented as follows. Section 2 gives preliminaries of some existing schemes considered in order to understand the proposed protocol. Section 3 explains the proposed symmetric-RSA (SYM-RSA) followed by its security analysis in Section 4. Section 5 contains the conclusion of the paper.

2. PRELIMINARIES Since the modifications of RSA are proposed using Diffie-Hellman (DH) key exchange protocol, thus DH and RSA techniques are introduced below.

2.1. Diffie-Hellman (DH) Protocol In (Diffie et al., 1976), Whitfield Diffie and Martin Hellman published an elementary article for secure exchange of a contributory common key between two remote participants over public channels. It does not require any prior information and is known to be the first public-key cryptosystem. In DH protocol, a finite multiplicative group with a generator g are publicly assumed, and two public messages are exchanged for negotiation of a secret key. Let A and B are two participants, who exchange the following two public messages, where A→ B: C means A sends message C to B: A→B: X = gx (mod P), where 1< x < P and x is a random secret chosen by A B→A: Y = gy (mod P), where 1< y < P and y is a random secret chosen by B The common contributory secret key K (say) is calculated by the participants independently as K =Yx(mod P) = (gy)x (mod P) =Xy (mod p) =(gx)y (mod P)

121

Modification of Traditional RSA into Symmetric-RSA Cryptosystems

This key K is secure, because the DHP (DH problem) and the underlying DLP (discrete logarithm problem) as given below are intractable in polynomial time for a large prime modulus P. DHP: Given public values (X = gx, Y = gy), the computation of K = gxy (mod P) DLP: Given either of the public values (X = gx or Y = gy), the computation of random secrets x or y This protocol has huge applications in designing other useful cryptosystems and some of them are Secure Socket Layer (SSL), Transport Layer Security (TLS), Internet Protocol Security (IPSec), Public Key Infrastructure (PKI), Digital Signature Standard (DSS), PGP (Pretty Good Privacy), SET (Secure Electronic Transaction) etc. Although the DH protocol has numerous applications, it is vulnerable to several attacks like man-in-the-middle, Denial-of-Service (DoS) etc.

2.2. RSA Algorithm The RSA was developed by Ron Rivest, Adi Shamir and Len Adleman in the year 1976 (Rivest et al., 1976), and is the most widely used public key algorithm in cryptography. It comprises following four steps: (1) (2) (3) (4)

Choose two large distinct prime numbers p and q Compute n = p × q, Euler function Ф(n) = (p-1) × (q-1) and delete p and q Select a number e such that1 < e < Ф(n) and relative prime to Ф(n) Compute multiplicative inverse d such that e × d ≡ 1mod Ф(n)

Encryption:

(1) Choose any k-bit plaintext m such that k ≤ log2 (n), where n is the public modulus. (2) Compute cipher-text c = me(mod n) Decryption: Compute plaintext m = cd(mod n)

3. MODIFIED ELGAMAL OVER INTEGERS PROPOSED MODIFICATION OF PUBLIC RSA TO SYMMETRIC RSA (SYM-RSA) As we know, Diffie- Hellman (DH) protocol supports negotiation of a secret key between two remote participants through the exchange of two public values, where pre-negotiation of a finite multiplicative large prime order group and a generator are necessary. Using these public values, both the parties calculate a common key independently. On the other hand, a user through RSA algorithm can generate public-private key pair for message encryption and other purposes. Now, we propose a technique to negotiate two prime numbers between two remote users using DH protocol and establish same publicprivate key pair by using RSA algorithm. Since two users have same RSA private key, the proposed scheme becomes a symmetric (SYM-RSA). In brief, SYM-RSA uses DH for secure negotiation of two prime numbers and then each user uses RSA for the generation of a common secret private key between them. Now, the details of the proposed scheme are described.

122

Modification of Traditional RSA into Symmetric-RSA Cryptosystems

3.1. Design of Proposed SYM-RSA It consist of two steps- (1) Use of DH key exchange protocol for secure negotiation of two prime numbers and (2) Use of RSA algorithm for generation of a common secret key. The step-1 is given below in Box 1. where two remote users A and B assume a finite multiplicative group with a generator and exchange four messages publicly. Note that two users A and B securely negotiate two prime numbers P1 and P2 using DH protocol, because two integers y, t < P, which are selected arbitrarily, cannot be expected to be considered by opponents although the public values RA1, RB1 are known. Each user then uses RSA with these two prime numbers and generates two common secret keys. In fact, the proposed scheme is framed in such a way that both the private- and public-key of the RSA including the common modulo become secret and only known to the users A and B, and hence, termed as SYM-RSA. The proposed scheme is explained through a flow-diagram as provided below in Box 2. As shown, the private- and public-key pair of RSA becomes secret, i.e., (e, n) and (d, n) are both secret and only known to the users A and B. As an application of the proposed SYM-RSA, user A may encrypt a message m using RSA algorithm using either of the two secret keys and send to user B. User B, on receiving the cipher-text, can decrypt easily using RSA and the remaining secret key as provided below: User A: Let the plaintext is m, so the cipher-text C = me mod n, here the secret key (e, n) is used for encryption (similarly, the secret (d, n) is also applicable). The cipher-text C is sent to B. User B: On receiving, user B decrypts C using the secret key (d, n) and gets plaintext as m = Cd mod n. Similarly, other applications of the traditional RSA can be implemented using our SYM-RSA cryptosystem. However, it may be noted that some new applications, which are not supported by existing RSA, could also be developed. For instance, simultaneous message security and integrity are easily possible in SYM-RSA, where the same is not supported in existing RSA. The diagrammatic view of the proposed signature process is given in Figure 1. and explained below. Let the plaintext be m, its simultaneous encryption and signature for integrity/authentication in one way can be done using SYM-RSA as follows: Encryption: C = me mod n Signature generation: Sig = hd mod n, where Sig stands for signature and h = H (m) for a hash function H. Final message: (C, Sig) Verification: m’ = C’d mod n, h’ = (Sig’) e mod n, where C’ and Sig’ are the received messages. Now, if h’ = H (m’), message security and integrity are verified, otherwise not. Note that if either of two secret keys is compromised, the SYM-RSA becomes a public-key RSA, which is exactly similar to the existing RSA.

123

Modification of Traditional RSA into Symmetric-RSA Cryptosystems

Box 1. Negotiation of two prime numbers using DH [Negotiate finite multiplicative group G (P) and a generator g] User A

User B

1) Randomly select x

1) Randomly select s

2) Compute and send to B: RA1 = g x mod P

→ 2) Compute and send to A: RB 1 = g s mod P

← y

t

3) Compute P1 = RB 1 mod P

3) Compute P2 = RA1 mod P

= gsy mod P

= gxt mod P (y and t are randomly selected such that both P1and P2become primes)

4) Compute and send to B: RA 2 = g y mod P

→ 4) Compute and send to A: t

RB 2 = g mod P

← x

s

5) Compute P2 = RB 2 mod P

5) Compute P1 = RA2 mod P

= gxt mod P

= gsy mod P

Box 2. Negotiation of two common secret keys using RSA (Prime numbers P1 and P2 are known) User A

User B

1) Calculate n= P1×P2

1) Calculate n = P1×P2

2) Ф (n) = (P1-1)(P2-1)

2)Ф(n) = (P1-1)(P2-1)

3) Select e, where gcd (e, Ф (n)) =1 [Steps 3 and 4 may be executed by B as well and similarly, the value of r be sent by B to A] 4) Compute r ≡P1×e mod P2 and send to B

r

→ −1

3) Compute e from r as e ≡ r × P1 5) Calculate d as e×d≡ 1mod Ф(n)

124

4) Calculate d as e×d≡ 1mod Ф (n)

mod P2

Modification of Traditional RSA into Symmetric-RSA Cryptosystems

Figure 1. Simultaneous message security and signature using SYM-RSA

4. IMPLEMENTATION This section shows the practical implementation of SYM-RSA using the mathematical example. However, for understanding purpose small numbers are taken, but practically it is considered as secure for < 512-bits. Example: Let us suppose two users A and B pre-negotiated a multiplicative group of prime order p=7853 with a generator g=3607. The required computations and message exchange are given below in Box 3. As shown, both e and d for the users A and B are unique and never be disclosed to others in our scheme.

5. SECURITY ANALYSIS As stated in the previous section, the proposed SYM-RSA uses DH and RSA public-key cryptosystems. The DH key exchange protocol is secured, because the underlying DHP (DH Problem) is computationally intractable. The DHP in the context of proposed SYM-RSA is restated below: As shown, two pair of messages (RA1 = gx mod P,RB2 = gt mod P) and (RA2 = gy mod P,RB1 = gs mod P) are publicly exchanged between two users A and B. Thus, a DHP, which exists corresponding to each pair, can be defined as the computation of P1 = gxt mod P (or P2 = gys mod P) by knowing the corresponding message pairs. This problem is not solvable in real time as no polynomial time algorithm is available. Also, as we know the hardness of DHP depends on the computation of either x or t (similarly either y or s), the solution of DHP is actually the solution of DLP (discrete logarithm problem) which is not tractable in polynomial time. Alternatively, the computation of P1 and/or P2 by an opponent on iteration over RA1 and RB1 for large prime modulo P is not feasible for having exact matching with secrets y and t, because these secrets are random and several such values exist for generating prime numbers from RA1 and RB1. Thus, the proposed prime numbers negotiation procedure is secure. On the other hand, the generation of two secret keys using RSA is also secured as explained below.

125

Modification of Traditional RSA into Symmetric-RSA Cryptosystems

Box 3. Negotiation of Two Prime Numbers Using DH User-A:

User-B:

1) Select a random number x=1239,

1) select a random number s= 976

2) Compute

RA1 = 36071239 mod 7853 = 546 RA1 = 546 to B

→ 2) Compute RB1 = 3607976mod7853 RB 1 = 7807 to A= 7807

← 3) Compute

P1 = 7807 779 mod 7853 = 3697

4) Compute

RA2 = 3607 779 mod 7853 = 3155

3) Compute

P2 = 546129 mod 7853 = 2749

Send RA2 = 3155 to B

→ 4) Compute

RB 2 = 3607129 mod 7853 = 543

Send RB 2 = 543 to A

← 5) Compute

P2 = 5431239 mod 7853 = 2749

5) Compute

P1 = 3155976 mod 7853 = 3697

Box 4. Negotiation of two common secret keys using RSA 1) Compute

n = 2749 × 3697 = 10163053

1) Compute

n = 2749 × 3697 = 10163053

2) Compute

Φ (n ) = (2749 − 1)(3697 − 1) = 10156608

2) Compute

Φ (n ) = (2749 − 1)(3697 − 1) = 10156608

3) Compute e= 991 such that gcd (991, 10156608)=1 4) Compute r ≡ 2749×991 mod 3697= 3267

5) Compute

126

d = 991−1 mod 10156608 = 9080479

Send r = 3267 to B

→ 3) Compute

e ≡ 3267 × 2749−1 mod 3697 = 991

4) Compute

d = 991−1 mod 10156608 = 9080479

Modification of Traditional RSA into Symmetric-RSA Cryptosystems

As we know, the existing RSA with a modulus greater than or equal to 512-bit number is secured, because no polynomial time algorithm in factorization of publicly known RSA modulus is available. Since the modulus n of the proposed SYS-RSA is not publicly known, our RSA modification in this respect is more secure than the existing one. Also, the computations of e, Ф (n) and d are hard as 1. Computation of e involves the solution of the congruence r = P1×e mod P2, where P1 and P2 are unknown (r is known). 2. Computation of Ф (n) involves on the availability of P1 and P2, however they are kept secret in our scheme. 3. Computation of d involves on the availability of e and Ф (n), which as shown in 1) and 2) are not computable. Thus, the proposed SYM-RSA as a whole is secured.

6. CONCLUSION This article presented a modification of existing RSA public-key cryptography and presented in the form of Symmetric-RSA (SYM-RSA) i.e. symmetric-key RSA is designed using the combination of DH and RSA algorithm. One of the application of proposed SYM-RSA is described i.e. message integrity. In addition, security analysis proves the security of proposed SYM-RSA in this paper as well. In future, the work will be extended and implement in real life scenario such as entity authentication.

REFERENCES Ambedkar, B. R., Gupta, A., Gautam, P., & Bedi, S. S. (2011, June). An Efficient Method to Factorize the RSA Public Key Encryption. Proceedings of the 2011 International Conference on Communication Systems and Network Technologies (CSNT) (pp. 108-111). IEEE. 10.1109/CSNT.2011.29 Diffie, W., & Hellman, M. (1976). New directions in cryptography. IEEE Transactions on Information Theory, 22(6), 644–654. doi:10.1109/TIT.1976.1055638 El-Kassar, A. N., Haraty, R. A., Awad, Y. A., & Debnath, N. C. (2005, November). Modified RSA in the Domains of Gaussian Integers and Polynomials Over Finite Fields (pp. 298–303). CAINE. Elkamchouchi, H., Elshenawy, K., & Shaban, H. (2002, November). Extended RSA cryptosystem and digital signature schemes in the domain of Gaussian integers. Proceedings of the 8th International Conference on Communication Systems (Vol. 1, pp. 91-95). 10.1109/ICCS.2002.1182444 Minni, R., Sultania, K., Mishra, S., & Vincent, D. R. (2013, July). An algorithm to enhance security in RSA. Proceedings of the 2013 Fourth International Conference on Computing, Communications and Networking Technologies (ICCCNT) (pp. 1-4). IEEE. 10.1109/ICCCNT.2013.6726517

127

Modification of Traditional RSA into Symmetric-RSA Cryptosystems

Mohit, P., & Biswas, G. P. (2015, March). Design of ElGamal PKC for encryption of large messages. Proceedings of the 2015 2nd International Conference on Computing for Sustainable Global Development (INDIACom) (pp. 699-703). IEEE. Mohit, P., & Biswas, G. P. (2016, March). Modification of Symmetric-Key DES into Efficient Asymmetric-Key DES using RSA. Proceedings of the Second International Conference on Information and Communication Technology for Competitive Strategies (p. 136). ACM. 10.1145/2905055.2905352 Peng, L., Hu, L., Lu, Y., Xu, J., & Huang, Z. (2016). Cryptanalysis of Dual RSA. Designs, Codes and Cryptography. Rivest, R. L., Shamir, A., & Adleman, L. (1983). A method for obtaining digital signatures and publickey cryptosystems. Communications of the ACM, 26(1), 96–99. doi:10.1145/357980.358017 Sun, H. M., Wu, M. E., Ting, W. C., & Hinek, M. J. (2007). Dual RSA and its security analysis. IEEE Transactions on Information Theory, 53(8), 2922–2933. doi:10.1109/TIT.2007.901248 Sun, H. M., Wu, M. E., Ting, W. C., & Hinek, M. J. (2007). Dual RSA and its security analysis. IEEE Transactions on Information Theory, 53(8), 2922–2933. doi:10.1109/TIT.2007.901248 Takagi, T. (1998, August). Fast RSA-type cryptosystem modulo p k q. Proceedings of the Annual International Cryptology Conference (pp. 318-326). Springer Berlin Heidelberg. 10.1007/BFb0055738 Thangavel, M., Varalakshmi, P., Murrali, M., & Nithya, K. (2015). An Enhanced and Secured RSA Key Generation Scheme (ESRKGS). Journal of Information Security and Applications, 20, 3–10. doi:10.1016/j.jisa.2014.10.004 Verkhovsky, B. (2011). Cubic Root extractors of Gaussian integers and their application in fast encryption for time-constrained secure communication. Int’l J. of Communications, Network and System. The Sciences, 4(4), 197.

This research was previously published in the International Journal of Business Data Communications and Networking (IJBDCN), 13(1); edited by Zoubir Mammeri; pages 66-73, copyright year 2017 by IGI Publishing (an imprint of IGI Global).

128

129

Chapter 8

Hybrid Approach of Modified AES Filali Mohamed Amine Djellali Liabes University, Algeria Gafour Abdelkader Djellali Liabes University, Algeria

ABSTRACT Advanced Encryption Standard is one of the most popular symmetric key encryption algorithms to many works, which have employed to implement modified AES. In this paper, the modification that has been proposed on AES algorithm that has been developed to decrease its time complexity on bulky data and increased security will be included using the image as input data. The modification proposed itself including alteration in the mix column and shift rows transformation of AES encryption algorithm, embedding confusion-diffusion. This work has been implemented on the most recent Xilinx Spartan FPGA.

INTRODUCTION It is important aspect to protect the confidential multimedia data from unauthorized access. Multimedia content can be text, audio, still images, animation and video. Such contents are protected by multimedia security method. Commonly, this is attained by techniques that are profoundly based on cryptography. These schemes facilitate communication security, piracy and shelter (Chang, 2014). Large size of images causes certain challenges for encryption. Normally a typical image has a very large size. Using traditional encryption algorithm will make encryption difficult for large volume of multimedia data (Chang et al, 2016). For the encryption of any multimedia data, we need such algorithms that require less computation because of large size of data. Symmetric-key algorithms are fewer computationally serious than any Asymmetric key algorithms. Typically, symmetric key algorithms are thousands of times sooner than those of the asymmetric algorithms. So, the better suitable method to encrypt the multimedia data is, to encrypt it with symmetric key encryption algorithms. DOI: 10.4018/978-1-7998-1763-5.ch008

Copyright © 2020, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Hybrid Approach of Modified AES

Our research is concerned with optimizing the existing standards of cryptography (AES) for the images and data encryption. It is also slanting towards exploiting the huge amount of data, in order to attain preferred speed.in this work proposes a modified version of AES algorithm and demonstrates that executions can accomplish superior and high throughput (Chang, et al., 2016).

STATE OF ART There are several modified in AES to improve speed the performance, increase the security and addition same complexity on algorithm steps. (Mohammad, et al., 2015) The reason development is appeared many different the implementation on software and hardware. Each implementation has need to modified AES according to the specific proposes. In (Xu, et al., 2013) a modified AES by used longer key length and data matrix. That extended the data matrix to eight row and variable number of column (6, 8, 12 and 16) the input data block (48, 64, 96 and 128), and extended the key length to (384,512, 768 and 1024). This paper not change the first and fourth stages (substitution byte, Add Round Key), third stage shift row change from shift third row to seventh row shifted left and four stages (mix column) change the static matrix 3x3 to new matrix 8x8, should be calculate inverse static matrix used in Mix column on GF (28). This modification is increase robustness and use a few times for encryption and decryption processing. In (Kaur, et al., 2014) modified the AES algorithm by reduce the calculation, computation overhead, and reduce the time encryption process. It replaces the mix column stage in AES algorithm into permutation stage (like the permutation table (IP) that used in DES algorithm) because the mix column is take large calculation time and that makes the encryption process are slow. The other stages in AES algorithm don’t change. Finally, modification used simple S-box for encryption and decryption to reduce the computation amount, the new S-box has some properties, simple generation and same S-box used for encryption and decryption (Rashidi, 2014).

The Advanced Encryption Standard (AES) The AES is a block cipher that is the standard version of Rijndael. It has a fixed block length of 128bits and variable key lengths. The number of internal rounds of the AES depends on the key size, which is 10, 12 and 14 for the key length 128,192 and 256 respectively. In our design, we consider the case of 128 bits for the key length and 10 rounds. Before the first round, the main key is added to the plaintext. Then, inrounds1–9, all four operations are performed to the state array. In the last round (10th round), the Mix-columns transformation is not used, which makes encryption and decryption symmetric (Daemen et Rijmen, 2001) (see Figure 1).

Sub Bytes Function performs a non-linear transformation independently on each byte of the input state. This transformation is performed by substituting each byte of the state with a value from substitution box (also termed as S-box). There are 16 parallel S-boxes each with eight inputs and eight outputs. The S-box operation is the only nonlinear transformation of the AES algorithm. It is an invertible operation and can be used for decryption processes too. 130

Hybrid Approach of Modified AES

Figure 1. The AES-128 encryption algorithm

Shift Row Shift Row operation is the cyclic shifting of each row of the state to the left. The shifting numbers are depended on the number of the row. The top row is not shifted and the last three rows are cyclically shifted over 1, 2, and 3 bytes, respectively

Mix Column This process is for mixing up of the bytes in each column separately during the forward process. The corresponding transformation during decryption is de noted InvMix-Columns and stands for the inverse mix column transformation (Ichikawa et al, 2000). The goal here is to further scramble up the 128-bit input block.

Key Addition In this operation, the round key is applied to the State by simple bit-by-bit XOR. Key-Addition is the same for the decryption process. Before the first round, a key addition layer is applied to the cipher data. This transformation is stated as the algorithm initial round key addition. The final round of the cipher is equal to the basic round with the Mix-Column step removed. A key expansion unit is defined in order to generate the appropriate key, for every round, from the initial key value. When all rounds of transformation are completed, a cipher data block with the same length as the plain data has been generated.

131

Hybrid Approach of Modified AES

Key Expansion The task of the key expansion module is to expand initial key for generating a series of Round Keys. In AES-128, this module generates a total of 10 Round Key of 16 bytes in order to be employed respectively in rounds of AES.

AES MODIFIED TECHNIQUE When we use the encryption algorithms for the security of complex multimedia data, computational overhead is so large that encryption becomes a hectic task (Wang et al, 2016). To overcome the problem of high calculation and computational overhead, we analyze the Advanced Encryption Standard (AES) and modify it, so as to reduce the calculation of algorithm and for improving the encryption performance security. So, we develop and implement a modified AES based Algorithm for all kinds of data. The basic aim of modifying the AES is to provide less computation and better security for data. The modified AES algorithm adjusts to provide better encryption speed.

Modified Aes Algorithm The present research modifies the AES algorithm (makes modification in Shift Row trans-formation and Mix-Columns) in order to make the AES algorithm more secure and faster. The four phases of the conventional AES constituting the round function are Sub Bytes, Shift-Rows, Mix-Columns and AddRoundkey. But the modified AES proposed is using Permutation in place of the Mix Column step. There are 10 rounds for full encryption. The four different stages that we use for Modified-AES Algorithm (see Figure 2) are: • • • •

Substitution bytes M-ShiftRows Permutation AddRoundKey

Modified Shift Row Transformation 1. When the round number is odd, the Shift Rows transformation operates on the rows of the state array (as normal Shift-Row transformation) (Abdulkarim, 2013) (see Figure 3); 2. When the round number is even, the Shift-Row transformation operates on the column of the state array; it shifts the bytes in each column by a certain offset in circle. The first column unchanged, the second column is shift cyclically shifted to the bottom by one byte, the third column is shifted cyclically to the bottom by two bytes and the last column is cyclically shifted to the bottom by three bytes.

132

Hybrid Approach of Modified AES

Figure 2. AES modified algorithm flow

Figure 3. Modified shift row transformation Case 2

Modified MixComls Permutation is widely used in cryptographic algorithms. Permutation operations are interesting and important from both the cryptographic and the architectural points of view. The inputs to the IP table consist of 64 bits. Modified-AES algorithm takes 128 bits as input. The functions Substitution Bytes and Shift Rows are also interpreted as 128 bits whereas the Permutation function takes 64 bits (Vandana, 2013). We divide the consequential bits of Shift Rows function into two parts of 64 bits, then take each part of 64 bits as input of permutation tables, and shift bits one by one according to that table. We fetch one bit from the source, and put it into the correct position in the destination. Each bit of a block is subject to initial permutation, which can be represented by the initial permutation (IP) table shown in Figure 4.

133

Hybrid Approach of Modified AES

Figure 4. IP table

This Modified AES algorithm takes 128-bit Sub bytes and Shift Rows operations also work data, there is a need to divide the sequential bits stage into 2 portions of 64 bits each and then part of 64 bits as an input of permutation tables according to IP table taken from DES algorithm. One bit from the source is fetched and then placed into position in target destination. Each bit is interpreted to IP table. After having completed the permutation of the 128 bits, we again repeat it for another 128-bit remaining operation of algorithm are executed (see Figure 5). Figure 5. IP-1 table

134

Hybrid Approach of Modified AES

For the full decryption of the Modified-AES algorithm, the transformation processes are, Inv-Bytesub, Inv-Shiftrows, Inv-Permutation, and the Addroundkey, which are performed in 10 rounds as it is in the encryption process.

RESULTS AND COMPARISONS The proposed modified AES encryption algorithm used to test and evaluate some bases image, which uses in encryption process based on software simulation (Chang et al., 2017). To test the algorithm, we take the different size of image and compare the calculated time of both the Modified-AES with Advanced Encryption Standard (AES). The result is shown in Table 1.

Digital Image Formats There are many types of digital image formats like.bmp, .gif, .jpg, .pict, .eps, and .png. This project can use any kind of format. The results obtain in this paper are mainly using .gif format only (Chang, et al, 2017) (see Figure 6). Table 1. Comparison of performance of AES with MAES Image Size (Pixel)

Image Size(kb)

AES-Time/ms

MAES-Time/ms

256*256

192

6.510

6.338

512*512

250

8.710

8.221

512*512

742

25.225

25.101

1024*1024

2.35 Mb

76.800

75.001

Figure 6. Encryption results time for different size of image

135

Hybrid Approach of Modified AES

Histograms of Encrypted Images We select grey-scale images (256×256) having different contents, and we calculate their histogram. One typical example among them is shown in Figure. 7. We can see that the histogram of the ciphered image is uniform and is significantly different from that of the original image. Therefore, it does not provide any indication to employ any statistical attack on the image under consideration. Moreover, there is no loss of image quality after performing the encryption/decryption steps.

Correlation of Two Adjacent Pixels Testing the correlation between two horizontally adjacent pixels, and two vertically adjacent pixels respectively, in a ciphered image. First, select n pairs of two adjacent pixels from an image randomly. Then the correlation coefficient of each pair uses the formula calculated in Figure 8. The results of the correlation coefficients for horizontal, vertical and diagonal adjacent pixels for the plain-image and its cipher-image are show in Table 2. The visual testing of the correlation distribution of two vertically adjacent pixels and the cipher image produced by the proposed scheme is shown in Figure 9. Figure 7. Histogram of original image and cipher image

136

Hybrid Approach of Modified AES

Figure 8. Equation for calculating image correlation

Table 2. Correlation coefficient of two adjacent pixels in original and encrypted image Direction

Plain Image

Ciphered Image

horizontal

0.9401

0.0093

vertical

0.9470

-0.0613

Diagonal

0.9122

-0.0081

Figure 9. Correlation of original image and cipher image

137

Hybrid Approach of Modified AES

FPGA IMPLEMENTATION The implementation of AES algorithm is performed on the Spartan 6 device, which is a member of Xilinx family. The VHDL code of the design is synthesized placed and routed using Xilinx ISE 14.1. Table 3 shows the results of the implementations of MAES in FPGA kit Spartan 6. It is clear from the table that increases the throughput. Because we optimized and modified the algorithm AES for speed, we can achieve 98.3 Gbps on a Xilinx device Spartan-6. Although our implementation needs more area, it has better area-speed utilization. For the sake of completeness, we have also compared results of our technique with the recent stateof-the-art results. For that purpose, we implemented our design on Xilinx Virtex-5 FPGA, since most of the recent work uses this FPGA as target device. We have compared the results of our technique and recent work against a number of area and performance parameter. Table 4 compares FPGA implementation results of our design to other selected designs. It is clear form Table 4 that our architecture Similarly, when we compare the best frequency results of existing technique, our technique gives comparable (in case of Virtex-5 implementation) or better results (in case of Virtex 5) while consuming much less device resources.

CONCLUSION The various modified AES algorithms proposed are more robust than the original AES as they have been tested on various parameters of security and the modified algorithm has proved to give better encryption results with reduced time complexity. A good encryption algorithm should resist various kinds of attacks say, known plain-text attack, and several other brute–force attacks. Table 3. Implementation results of our AES modified and AES classic Encryption Algorithm

Latency

Frequency (MHZ)

Throughput (Gbps)

Area %

M-AES

3.354

886.64

98.3

49.4

AES

1.538

650.364

68.1

64.2

Table 4. Comparison between our implementations and other previous implementations Design

Type FPGA

Frequency Mhz

Throughput(Gb/s)

Wang et Ha

Virtex5

312.80

70.9

Soliman et al.

Virtex5

50.20

73.37

Van et al.

Virtex5

619

64.7

Qu et al.

Virtex5

258.5

18.5

Rahimunnisa et al.

Virtex5

/

20.3

Our design

Virtex5

886.64

88.1

138

Hybrid Approach of Modified AES

Future Work We will perform the security analysis of MAES, to see how; much more or less secure it is, as compare to the different AES implementation. Moreover, we will compare it with other Modified versions of AES, work done by other researchers.

REFERENCES Anwar, H., Daneshtalab, M., Ebrahimi, M., Plosila, J., & Tenhunen, H. (2013, December). FPGA implementation of AES-based crypto processor. In Proceedings of the 2013 IEEE 20th International Conference on Electronics, Circuits, and Systems (ICECS) (pp. 369-372). IEEE. Chang, V., Amin, R., Islam, S. H., Vijayakumar, P., & Khan, M. K. (2017). A robust and efficient bilinear pairing based mutual authentication and session key verification over insecure communication. Multimedia Tools and Applications. Chang, V., Amin, R., Kumar, N., Biswas, G. P., & Iqbal, R. (2016). A light weight authentication protocol for IoT-enabled devices in distributed Cloud Computing environment. Future Generation Computer Systems. Chang, V., Campbell, J., & Et Hosseinian-Far, A. (2016). Philosophising data: a critical reflection on the ‘hidden’ issues. In Big Data: Concepts, Methodologies, Tools, And Applications (pp. 302–313). Hershey, PA: IGI GLOBAL. Chang, V., Kuo, Y. H., & Ramachandran, M. (2016). Cloud computing adoption framework: A security framework for business clouds. Future Generation Computer Systems, 57, 24–41. doi:10.1016/j. future.2015.09.031 Chang, V., & Ramachandran, M. (2014, January). A proposed case for the cloud software engineering in security. In Proceedings of the International Workshop on Emerging Software as a Service and Analytics, ESaaSA 2014-In Conjunction with CLOSER 2014 (pp. 71-79). Daemen, J., & Rijmen, V. (2001, November). Advanced Encryption Standard. National Institute for Standards and Technology (NIST). Retrieved from http://csrc.nist.gov/publictions/fips/fips-197 Hoang, T. (2012, February). An efficient FPGA implementation of the Advanced Encryption Standard algorithm. In Proceedings of the 2012 IEEE RIVF International Conference on Computing and Communication Technologies, Research, Innovation, and Vision for the Future (RIVF), (pp. 1-4). IEEE. 10.1109/rivf.2012.6169845 Hodjat, A., & Verbauwhede, I. (2004, April). A 21.54 Gbits/s fully pipelined AES processor on FPGA. In Proceedings of the 12th Annual IEEE Symposium on Field-Programmable Custom Computing Machines FCCM ’04 (pp. 308-309). IEEE. Huang, C. W., Chang, C. J., Lin, M. Y., & Tai, H. Y. (2007, October). Compact FPGA implementation of 32-bits AES algorithm using Block RAM. In Proceedings of the TENCON 2007 IEEE Region 10 Conference (pp. 1-4). IEEE.

139

Hybrid Approach of Modified AES

Ichikawa, T., Kasuya, T., & Matsui, M. (2000, April). Hardware Evaluation of the AES Finalists. In Proceedings of the AES candidate conference (pp. 279-285). Järvinen, K. U., Tommiska, M. T., & Skyttä, J. O. (2003, February). A fully pipelined memoryless 17.8 Gbps AES-128 encryptor. In Proceedings of the 2003 ACM/SIGDA eleventh international symposium on Field programmable gate arrays (pp. 207-215). ACM. Kaur, A., Bhardwaj, P., & Kumar, N. (2013). Fpga implementation of efficient hardware for the Advanced Encryption Standard. International Journal of Innovative Technology and Exploring Engineering, 2(3), 186–189. Koradia, V. (n. d.). Modification in Advanced Encryption Standard. Journal of Information, Knowledge and Research in Computer Engineering, 2(2), 358. Mohammad, O. K. J., Abbas, S., El-Horbaty, E. S. M., & Salem, A. B. M. (2015). Innovative method for enhancing key generation and management in the aes-algorithm. arXiv:1504.03406 National Institute of Standards and Technology (U.S.). (n. d.). Advanced Encryption Standard. Retrieved from http://csrc.nist.gov/publication/drafts/dfips-AES.pdf Qin, H., Sasao, T., & Iguchi, Y. (2006). A design of AES encryption circuit with 128-bit keys using look-up table ring on FPGA. IEICE Transactions on Information and Systems, 89(3), 1139–1147. doi:10.1093/ietisy/e89-d.3.1139 Qu, S., Shou, G., Hu, Y., Guo, Z., & Qian, Z. (2009, May). High throughput, pipelined implementation of AES on FPGA. In Proceedings of the International Symposium on Information Engineering and Electronic Commerce IEEC ’09 (pp. 542-545). IEEE. 10.1109/IEEC.2009.120 Rahimunnisa, K., Karthigaikumar, P., Rasheed, S., Jayakumar, J., & SureshKumar, S. (2014). FPGA implementation of AES algorithm for high throughput using folded parallel architecture. Security and Communication Networks, 7(11), 2225–2236. doi:10.1002ec.651 Rahimunnisa, K., Karthigaikumar, P., Rasheed, S., Jayakumar, J., & SureshKumar, S. (2014). FPGA implementation of AES algorithm for high throughput using folded parallel architecture. Security and Communication Networks, 7(11), 2225–2236. doi:10.1002ec.651 Rashidi, B., & Rashidi, B. (2013). Implementation of an optimized and pipelined combinational logic rijndael S-Box on FPGA. International Journal of Computer Network and Information Security, 5(1), 41–48. doi:10.5815/ijcnis.2013.01.05 Saggese, G. P., Mazzeo, A., Mazzocca, N., & Strollo, A. G. (2003, September). An FPGA-based performance analysis of the unrolling, tiling, and pipelining of the AES algorithm. In FPL (pp. 292-302). Saggese, G. P., Mazzeo, A., Mazzocca, N., & Strollo, A. G. (2003, September). An FPGA-based performance analysis of the unrolling, tiling, and pipelining of the AES algorithm. In FPL (pp. 292-302). Saqib, N. A., Rodríguez-Henríquez, F., & Díaz-Pérez, A. (2003, September). AES algorithm implementation-an efficient approach for sequential and pipeline architectures. In Proceedings of the Fourth Mexican International Conference on Computer Science ENC ‘03 (pp. 126-130). IEEE. 10.1109/ ENC.2003.1232885

140

Hybrid Approach of Modified AES

Shtewi, A. A., Hasan, B. E. M., & Hegazy, A. E. F. A. (2010, February). An Efficient MAES Adapted for Image Cryptosystems. International Journal of Computer science and Network Security, 10(2). Soliman, M. I., & Abozaid, G. Y. (2011). FPGA implementation and performance evaluation of a high throughput crypto coprocessor. Journal of Parallel and Distributed Computing, 71(8), 1075–1084. doi:10.1016/j.jpdc.2011.04.006 Soliman, M. I., & Abozaid, G. Y. (2011). FPGA implementation and performance evaluation of a high throughput crypto coprocessor. Journal of Parallel and Distributed Computing, 71(8), 1075–1084. doi:10.1016/j.jpdc.2011.04.006 Van Dyken, J., & Delgado-Frias, J. G. (2010). FPGA schemes for minimizing the power-throughput trade-off in executing the Advanced Encryption Standard algorithm. Journal of Systems Architecture, 56(2), 116–123. doi:10.1016/j.sysarc.2009.12.001 Verbauwhede, I., Schaumont, P., & Kuo, H. (2003). Design and performance testing of a 2.29-GB/s Rijndael processor. IEEE Journal of Solid-State Circuits, 38(3), 569–572. doi:10.1109/JSSC.2002.808300 Wang, Y., & Ha, Y. (2013). FPGA-based 40.9-Gbits/s masked AES with area optimization for storage area network. IEEE Transactions on Circuits and Systems, 60(1), 36–40. doi:10.1109/TCSII.2012.2234891 Wang, Z., Cao, C., Yang, N., & Chang, V. (2016). ABE with improved auxiliary input for big data security. Journal of Computer and System Sciences. Xu, P., Jin, H., Wu, Q., & Wang, W. (2013). Public-key encryption with fuzzy keyword search: A provably secure scheme under keyword guessing attack. IEEE Transactions on Computers, 62(11), 2266–2277. doi:10.1109/TC.2012.215

This research was previously published in the International Journal of Organizational and Collective Intelligence (IJOCI), 7(4); edited by Victor Chang; pages 83-93, copyright year 2017 by IGI Publishing (an imprint of IGI Global).

141

142

Chapter 9

Cryptographic Algorithms for Next Generation Wireless Networks Security Vishnu Suryavanshi GHRCE Nagpur, India G. C. Manna BSNL, India

ABSTRACT At present a majority of computer and telecommunication systems requires data security when data is transmitted the over next generation network. Data that is transient over an unsecured Next Generation wireless network is always susceptible to being intercepted by anyone within the range of the wireless signal. Hence providing secure communication to keep the user’s information and devices safe when connected wirelessly has become one of the major concerns. Quantum cryptography algorithm provides a solution towards absolute communication security over the next generation network by encoding information as polarized photons, which can be sent through the air security issues and services using cryptographic algorithm explained in this chapter.

INTRODUCTION Does increased security provide comfort to paranoid people? Or does security provide some very basic protections that we are naive to believe that we don’t need? During this time when the Internet provides essential communication between tens of millions of people and is being increasingly used as a tool for commerce, security becomes a tremendously important issue to deal with. There are many aspects to security and many applications, ranging from secure commerce and payments to private communications and protecting passwords. One essential aspect for secure communications in next generation wireless network is that of cryptography, which the focus of this chapter is. DOI: 10.4018/978-1-7998-1763-5.ch009

Copyright © 2020, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Cryptographic Algorithms for Next Generation Wireless Networks Security

Security in computer world determines the ability of the system to manage, protect and distribute sensitive information (Abdel-Karim R. Al Tamimi 2006). The most attractive and fast growing network is 802.11 in wireless networks. In 1997IEEE 802.11 introduced standards for wireless local network (WLAN) communication, some of these standards are: • • •

Using the 2.4 GHz radio spectrum and 11 Mbps max data rate is 802.11b. Using the 5 GHz radio spectrum and 54 Mbps max data rate is 802.11a. Using the 2.4 GHz radio spectrum and 54 Mbps max data rate is 802.11g.

Wireless Robust Security Network is 802.11i (Quality of service). It is used in quality of service for traffic prioritization to give delay sensitive application such as multimedia and voice communication priority(SANS, 2005).Next generation wireless technology 3G, 4G and more has been gaining rapid popularity in recent years. They have ubiquitous wireless communications and services as Integration of multi-networks is using IP technology; similar technology to the wired Internet where users are freed from their local networks, not just IP end-to-end but over-the-air packet switching, high bandwidth / high-speed wireless and highly compatible with wired network infrastructures like ATM, IP. These technologies are facing security problems in the software products used to access the vast Internet, operating systems, www browsers and e-mail programs(Chandra, et al., 2008).For secure data transformation cryptographic algorithms plays a key role. A cryptographic technique provides three forms of security namely confidentiality, data integrity and authentication. Confidentiality refers to protection of information from unauthorized access (Daemean & Rijmen, 1999). Information has not been manipulated in any unauthorized way is ensured by data integrity. Authentication can be explained in two groups as entity authentication and message authentication. Detecting any modifications to the message provides message authentication. Entity authentication assures the receiver of a message, about both the identity of the sender and his active participation (Kumar & Purohit, 2010) Need of a standard depends on the ease of use and level of security which it provides. Here, the distinction between wireless usage and security standards show that the security is not maintained well up to with the growth past of end user’s usage. The hackers monitor and even change the integrity of transmitted data in current wireless technology. Lack of rigid security standards has caused companies to invest millions on securing their wireless networks. Securing Next Generation Wireless Networks is an extremely challenging and interesting area of research. Unprotected wireless networks are vulnerable to several security attacks including eavesdropping and jamming that have no counterpart in wired networks. Moreover, many wireless devices are resource limited, which makes it challenging to implement security protocols and mechanisms. The main objective of this chapter to study and analyze use of Cryptographic Algorithms for Next Generation wireless networks Security in terms confidentiality, Confidentiality, Integrity, Availability, Anti-virus, anti-spyware software, firewall, Authentication, Access control, and Cryptanalysis.

143

Cryptographic Algorithms for Next Generation Wireless Networks Security

BACKGROUND Information is an important asset and resource for business and needs to be protected like any other asset. The protection of information is usually known as information security. A basic and classical model of security objectives include the so called CIA triad which stands for confidentiality, integrity and availability. In this chapter we tried to notify security issues and services and use of cryptographic algorithms.

1. SECURITY ISSUES AND SERVICES Signal fading, mobility, data rate enhancements, minimizing size and cost, user security and (Quality of service) QoS are the key challenges in wireless networks (Kumar & Jain, 2012).Handheld devices which are used in embedded application have not generally been viewed as posing security threats, their increased computing power and the ease with which they can access networks and exchange data with other handheld devices introduce new security risks to an agency’s computing environment. This section describes how the security requirements for confidentiality, integrity, authenticity, and availability for handheld device computing environments can be threatened. In a sizable geographic area, Wireless Mesh Networks (WMNs) presents a good solution to provide wireless Internet connectivity (Siddiqui & Seon, 2007).

2. SECURITY ISSUES 2.1 Loss of Confidentiality Confidentiality assures only the knowing recipients is accessible to the information transmitted across the network. On the handheld device, the storage module, or the PC or while being sent over one of the Bluetooth, 802.11, IR, USB, or serial communication ports; confidentiality of information can be compromised. Moreover, most handheld devices are shipped with connectivity that is enabled by default. These default configurations are typically not in the most secure setting and should be changed to match the agency’s security policy before being used. Confidentiality is the privacy of a useful thing. Specifically, confidentiality can be defined as which people, under what conditions are authorized to access a useful thing. The confidentiality of this information is extremely important because the subjection of this information could bring embarrassment and heavy penalties. Confidentiality can be achieved through strong asymmetric cryptographic solutions in wired networks (Anjum & Salil, 2009)

2.2 Loss of Integrity Information secrecy, data integrity and resource availability of users are to be provided by security services. To prevent improper modification of data and resource availability preventing denial of services is data integrity (Liang & Wang, 2004).

144

Cryptographic Algorithms for Next Generation Wireless Networks Security

The integrity of the information on the handheld device and of the handheld device hardware, applications, and underlying operating system are also security concerns. Information stored on, and software and hardware used by, the handheld device must be protected from unauthorized, unanticipated, or unintentional modification. Information integrity requires that a third party be able to verify that the content of a message has not been changed in transit and that the origin or the receipt of a specific message be verifiable by a third party Integrity is evaluated by two primary properties. First, there is the notion that a useful thing should be trusted; that is, there is an expectation that a useful thing will only be modified in appropriate ways by appropriate people. For example, a hospital patient’s allergy information is stored in a database. The doctor should be able to trust that the allergy information is correct and up-to-date. If data is damaged or incorrectly altered by authorized or unauthorized personnel then you must consider how important it is that the data be restored to a trustworthy state with minimum loss; which is the second part of integrity. For example, suppose a nurse who is authorized to view and update a patient’s allergy information is upset with his/her employer and wants to disrupt a patient’s data to make the hospital look bad. How important is it that the hospital be able to catch this error and trace it back to the person(s) who caused it? Data integrity is guaranteed because required keys may be generated during the authentication process for data encryption and message authentication.

2.3 Loss of Availability In the modeling and design of fault tolerant wireless systems, availability for wireless mobile systems has presented great challenges. There are high expectations from customer for level of availability and performance from wireless communication system as with the rapid growth of wireless communication services (Kishor, Trivedi, Ma & Dharmaraja, 2003). Authorized person, entity, or device can access a useful thing represented by availability. For example, an organization has a system which provides authentication services for critical systems, applications, and devices on campus. An interruption in this service could mean the inability for customers to access computing resources and staff to access the resources they need to perform critical tasks. Therefore a loss of the service could quickly translate into a large financial loss in lost employee time and potential customer loss due to inaccessibility of resources. Because of this, the availability of this authentication system would be considered ‘High’. The outage-and-recovery of its supporting functional units can affect the performance and availability of a wireless system.

2.4 Anti-Virus and Anti-Spyware Software, and a Firewall Now a day’s antivirus solution became a normal component of computer system. Just like other components or services of computer system antivirus software can be targeted. Anti-virus software is the most cumbersome implementation. It has to pass with hundreds of file types and formats like executables, documents, compressed archives, executable packers and media files. Such formats are quite complex. Hence to implement such software on these formats is extremely difficult (Feng, 2008). Install anti-virus and anti-spyware software, and keep them up-to-date.

145

Cryptographic Algorithms for Next Generation Wireless Networks Security

For any program that tracks user’s online activities and secretly transmits information to a third party is a spyware. Annoying interruptions like pop-ups ads to security breaches and loss of intellectual property are the effects of spyware (Webroot Software Inc., 2004) Unauthorized access to or from a network is prevented by firewall which is hardware or software system. It can be implemented in both hardware and software, or a combination of both. Firewalls are used to prevent unauthorized Internet users from accessing private networks connected to the Internet. All data entering or leaving the Intranet pass through the firewall, which examines each packet and blocks those that do not meet the specified security criteria (Okumoku-Evroro & Oniovosa, 2005) If your firewall was shipped in the “off” mode, turn it on.

3. SECURITY SERVICES 3.1 Authentication Entity authentication (or “peer entity authentication” as it is referred to in ISO 7498-2) provides corroboration to one entity that another entity is as claimed. This service provides confidence, at the time of use only, that an entity is not attempting to impersonate another entity or that the current communications session is an unauthorized replay of a previous connection. To protect credit card transactions on the Internet the Secure Electronic Transaction is used. IBM, Microsoft, Netscape, RSA, Terisa and Verisign these companies are collaborated in the development of SET. This system is designed for Wired networks and does not meet all the challenges of wireless network. Multifactor Authentication techniques can be used to provide secure web transactions using cell phones. This multifactor technique is based on TIC’s and SMS confirmation (Tiwari & Sudip, 2007) Authentication is a process which a user gains the right to identify him or her. Passwords, biometric techniques, smart cards, certificates, etc. are the key techniques to authenticate a user. These techniques namely come under Multifactor Authentication techniques. Usually within one institute, a user may have a single identity; however, if a user has rights to identify himself in several different organizations or systems, more than one identity from a person may cause problems. There are four major scenarios based on different degrees of trust: 1. The right of an individual to self-determination as to the degree to which personal information will be shared among other individuals or organizations to control the collection, storage, and distribution of personal or organizational information. 2. The right of an individual to self-determination as to the degree to which the individual is willing to share with others information about himself that may be compromised by unauthorized exchange of such information among other individuals or organizations 3. The right of individuals and organizations to control the collection, storage, and distribution of their information or information about themselves. 4. The right of individuals to control or influence what information related to them may be collected and stored and by whom and to whom that information may be disclosed.

146

Cryptographic Algorithms for Next Generation Wireless Networks Security

Authentication mechanisms differ in the assurances they provide: 1. Data was generated by the Principal at some point in the past. 2. The Principal was present when the data was sent. 3. The data received was freshly generated by the Principal. Mechanisms also differ in the number of verifiers: 1. Support for single verifier per message. 2. Support for multiple verifiers. Whether the mechanism supports the ability of the verifier to prove to a third party that the message originated with the Principal. We divide the authentication policy into three major categories: 1. Personal/system; 2. Internet; 3. Network authentication.

3.2 Access Control To control the flow of information between subject and object where subject is always an active entity while object is a passive entity this mechanism is called access control. Access control is a three-step process which includes identification, authentication and authorization. There are three access control modes which have their own merits and demerits. These are • • •

Discretionary Access Control (DAC). Mandatory Access Control (MAC). Role Based Access Control (RBAC).

The first step in any access control solution is identification or authentication. Authentication are often discussed in terms of “factors” of proof, such as a PIN, a smart card and a fingerprint. Access Control Techniques and Technologies are Rule Based Access Control, Menu Based Access Control (Vinay, 2007). Over the years security practitioners have developed a number of abstractions in dealing with access control. Protection of objects is the crucial requirement, which in turn facilitates protection of other resources controlled via the computer system. Access control provides a secure solution for web services (WS). To find syntactic and semantic errors administrators of WS can specify access control policies and validate them (Yague & Javier, 2005). •

For the transmission of data the users belonging to the same multicast session form a Data Group (DG). One DG contains the users that can access to a particular resource. According to access privilege the users are also divided into non-overlapping Service Groups (SG).

147

Cryptographic Algorithms for Next Generation Wireless Networks Security

• •

Access control on manipulation of resources via “Hyper Text Transfer Protocol” Adam’s user agent attempts the reading, writing, or deletion of an information resource identified by a Universal Research Locator (URL); Adam’s user agent attempts the use of a processing resource to execute programs.

3.3 Data Confidentiality Robust Security Network Association (RSNA) provides two data confidentiality protocols, called the Temporal Key Integrity Protocol (TKIP) and the Counter-mode/CBC-MAC Protocol (CCMP). Confidentiality protection schemes are: • • •

Data Privacy: The data produced by each sensor node should be only known to itself. Data Confidentiality: In addition to data privacy, partially or fully aggregated data should only be known by the sink. Efficiency: After the confidentiality protection schemes are introduced, the system overhead should be kept as small as possible (Taiming, 2001).

Guidelines have been issued in a number of specific areas to help to protect the confidentiality of personal data held in the department. 1. The first stage in establishing policies and procedures to ensure the protection of personal data is to know: a. What data is held? b. Where it is held, and c. What are the consequences would be when data is lost or stolen. With that in mind, as a first step identifying the types of personal data held within the department, identifying and listing all information repositories holding personal data and their location should conduct an audit. The storage, handling and protection of this data should be included in the Department’s risk register these are associated with risks. The security measures in place are appropriate and proportionate to the data being held can be established by department. 2. All data centers and server rooms used to host hardware and software on which personal data is stored should be restricted to access. This can be done using entail swipe card and/or PIN technology to the room(s) in question – such a system should record when, where and by whom the room was accessed. Such access records and procedures should be reviewed by management frequently. 3. Those computer systems which are no longer in active use and which contain personal data should be removed. 4. Passwords used to access PCs, applications, databases, etc. should be of sufficient strength. A password should include numbers, symbols, upper and lowercase letters. If possible, password length should be around 12 to 14 characters but at the very minimum 6 to 8 characters. Repetition, dictionary words, letter or number sequences, usernames, or biographical information like names or dates must be avoided as Passwords. They should be changed on a regular basis (CMOD, 2008). 148

Cryptographic Algorithms for Next Generation Wireless Networks Security

It defines four types of data confidentiality service:

3.4 Connection Confidentiality This service provides for the confidentiality of all user data transferred using a connection. • • •

Connectionless Confidentiality: This service provides for the confidentiality of all user data transferred in a single connectionless data unit (i.e. a packet). Selective Field Confidentiality: This service provides for the confidentiality of selected fields within user data transferred in either a connection or a single connectionless data unit. Traffic Flow Confidentiality: This service provides for the confidentiality of information which might be derived from observation of traffic flows.

4. CRYPTOLOGY AND ITS CLASSIFICATION Cryptology has two main branches cryptography and cryptanalysis.

4.1 Cryptanalysis The mathematical science that deals with analysis of a cryptographic system to gain knowledge needed to break or circumvent the protection that the system is designed to provide. Attacks, in the context of network security, can be classified in two main classes, active and passive. The many known attacks against WEP can be categorized into different groups according to their goals: Cryptosystems come in 3 kinds: 1. Those that have been broken (most). 2. Those that have not yet been analyzed (because they are new and not yet widely used). 3. Those that have been analyzed but not broken. (RSA, Discretelog cryptosystem, AES). Most common ways to turn cipher text into plaintext: 1. Steal/purchase/bribe to get key 2. Exploit sloppy implementation/protocol problems (hacking/cracking) examples someone used spouse’s name as key, someone sent key along with message. The main goal of a cryptanalyst is to obtain maximum information about the plaintext (original data). a. Message Decryption: Allows the attacker to obtain the plaintext corresponding to the ciphertexts of messages intercepted in the network. b. Message Injection: Allows the attacker to actively generate new valid messages and send them to stations associated with the network. c. Key Recovery: Is the process of obtaining the pre-shared WEP key. This is the most interesting attack type, because successfully executed it allows the attacker full access to the network

149

Cryptographic Algorithms for Next Generation Wireless Networks Security

5. LINEAR CRYPTANALYSIS High probability occurrences of linear expressions involving plaintext bits, “ciphertext” bits, and sub-key bits will be the advantage of linear cryptanalysis. It is a known plaintext attack: that is, it is premised on the attacker having information on a set of plaintexts and the corresponding ciphertexts. However, the attacker has no way to select which plaintexts (and corresponding ciphertexts) are available. The attacker has knowledge of a random set of plaintexts and the corresponding ciphertexts in many applications and scenarios. The basic idea is to approximate the operation of a portion of the cipher with an expression that is linear where the linearity refers to a mod-2 bit-wise operation (i.e., exclusive-OR denoted by “⊕”). Such an expression is of the form: Xi1⊕Xi2⊕… ⊕Xin⊕Yj⊕Yj1⊕Yj2⊕….Yjn = u v

(1)

where Xi represents the i-th bit of the input X = [X1, X2, ...] and Yj represents the j-th bit of the output Y = [Y1, Y2, ...]. This equation is representing the exclusive-OR “sum” of u input bits and v output bits.

6. DIFFERENTIAL CRYPTANALYSIS High probability of certain occurrences of plaintext differences and differences into the last round of the cipher exploits the Differential cryptanalysis. For example, consider a system with input X = [X1 X2 ... Xn] and output Y = [Y1 Y2 ... Yn]. Let two inputs to the system be X′ and X″ with the corresponding outputs Y′ and Y″, respectively. The input difference is given by ∆X = X′ ⊕ X″ where “⊕” represents a bit-wise exclusive-OR of then-bit vectors and, hence, [∆X] = [∆X1 ∆X2 …∆Xn] where ΔXi=X′i⊕ X′′I with Xi′ and Xi″ representing the i-th bit of X′ and X″, respectively. Similarly, ∆Y = Y′ ⊕ Y″ is the output difference and [∆Y] = [∆Y1 ∆Y2…. ∆Yn] where ∆Yi =Y′⊕Y ′′. In an ideally randomizing cipher, the probability that a particular output difference ∆Y occurs given a particular input difference ∆X is ½n where n is the number of bits of X. A particular ∆Y occurs given a particular input difference ∆X with a very high probability pD (i.e., much greater than (1/2n) seeks to exploit a scenario in Differential cryptanalysis The pair (∆X, ∆Y) is referred to as a differential. The Differential cryptanalysis which is a chosen plaintext attack, the attacker is able to select inputs and examine outputs in an attempt to derive the key. For differential cryptanalysis, the attacker will select pairs of inputs, X′ and X″, to satisfy a particular ∆X, knowing that for that ∆X value, a particular ∆Y value occurs with high probability (Heys, 2001)

150

Cryptographic Algorithms for Next Generation Wireless Networks Security

7. CRYPTOGRAPHY Cryptography is the study of how to design algorithms that provide confidentiality, authenticity, integrity and other security related services for data transmitted in insecure communication environments. Confidentiality protects data from leaking to unauthorized users. Authenticity provides assurance regarding the identity of a communicating party, which protects against impersonation. Integrity protects data against being modified (or at least enables modifications to be detected). For the encryption and decryption of algorithm if we are using same key (i.e. for encryption and decryption) then it is called symmetric key cryptography or it is also called one key algorithm, where as in asymmetric or public key cryptography requires two keys one is used to encrypt the plaintext and other to decrypt the cipher text. One of these keys is published or public and the other is kept private. Cryptography is classified into four categories.

7.1 Symmetric Key Cryptography Symmetric-key encryption is that both the sender and the receiver of an encrypted message have a common secret key k (see Figure 2.). In order to encrypt the message m, also referred to as plaintext, the sender uses the function E together with the key c = Ek(m) D and secret key k (m = message, c = encrypted message). One should assume that E and D are known to the public in that way an encryption scheme is designed, and obtaining the message m from ciphertext c merely depends on the secret key k (principle of Kerckhoff). In practice, the principle of Kerckhoff is not always used. That means that the encryption scheme is kept secret. There are two reasons for this: one can adopt an even higher security through this additional secrecy, to protect a system not only against cryptographic attacks but also against attacks on

Figure 1. Evolution of cryptography

151

Cryptographic Algorithms for Next Generation Wireless Networks Security

Figure 2. Symmetric-key encryption system with encryption function E, decryption function

the hardware. Secondly, the use of a weak and inadequately examined algorithm is concealed through secrecy. Based on symmetric keys a new robust cryptography algorithm to increase security and prevent from unauthorized access to the contents of encrypted files is developed. It depends on structure of files; creation method of keys and resultantly the secret file cryptography using each of them are the key factors (Mohammad, 2013).

7.2 Public Key Cryptography From the last 300-400 years Public-key cryptography (PKC) is the most significant new development in cryptography. Modern PKC was first described publicly by Stanford University professor Martin Hellman and graduate student Whitfield Diffie in 1976. Public-key introduces another concept involving key pairs: one for encrypting, the other for decrypting. This concept is very clever and attractive, and provides a great deal of advantages over symmetric-key: • • •

Simplified key distribution; Digital Signature; Long-term encryption.

7.3 Un-Keyed Cryptography Un-keyed cryptography study gives details of hash function. Hash functions are very important primitives in cryptography. Hash functions can be used to protect the authenticity of information and to improve digital signature schemes. The protection of the authenticity of information includes two aspects: • •

152

The protection of the originator of the information, or in ISO terminology data origin authentication, The fact that the information has not been modified or in ISO terminology the integrity of the information (Preneel, 2003).

Cryptographic Algorithms for Next Generation Wireless Networks Security

Hash function in cryptography is classified in two types: • •

Un-keyed hash functions, and Keyed hash functions.

Cryptography un-keyed hash functions should satisfy preimage resistance and second preimage resistance and collision resistance. Preimage resistance means that, given an output it is infeasible to obtain an input which produces the output. Second preimage resistance is that, given an input it is infeasible to obtain an input which produces the same output as the given input. Lastly it is infeasible to obtain two different inputs which produce the same output as given input is called collision resistance.

8. CRYPTOGRAPHIC PROTOCOLS A security protocol (cryptographic protocol or encryption protocol) is that performs a security-related function and applies cryptographic methods, often as sequences of cryptographic primitives. Security protocols are small programs that aim at securing communications over a public network, like Internet. A variety of such protocols has emerged and is seeing increasing use. SSL/TLS, SSH, and IPsec are used in internet traffic (Como, Cortier, & Zalinescu, 2009). How the algorithms should be used describes a protocol. A sufficiently detailed protocol includes details about data structures and representations, at which point it can be used to implement multiple, interoperable versions of a program. Cryptographic protocols are widely used for secure application-level data transport, some of these aspects: • • • • • • •

Key agreement or establishment; Entity authentication; Symmetric encryption and message authentication material construction; Secured application-level data transport; Non-repudiation methods; Secret sharing methods; Secure multi-party computation.

9. STREAM CIPHERS AND BLOCK CIPHERS Secret key cryptography schemes are generally categorized as stream and block cipher.

9.1 Stream Cipher Stream ciphers, which belong to the symmetric encryption techniques. Design and analysis of stream cipher systems as well as the most well-known encryption systems are introduced. When a block cipher is used, a long message m is divided into blocks m =m0,m1, . . ., mN−1 of the same length. Here the blocks have usually a length of n = 64, 128 or 256 bits, depending on the processing length n of the block cipher.

153

Cryptographic Algorithms for Next Generation Wireless Networks Security

When stream ciphers are used, the message to be encrypted m is also divided into blocks. Here, however, only short blocks of length n occur. In this case we do not speak of a division into blocks, but into symbols. Usually, n = 1 or n = 8 bit. The encryption of the single symbols mt is carried out through a state dependent unit. Stream ciphers are slower than block and the Transmission error in one cipher text block have effect on other block such that if a bit lost or a altered during transmission the error affect the nth character and cipher re-synchronous itself after n correct cipher text characters as well as not suitable in the software. In synchronous stream cipher if a cipher text character is lost during transmission the sender and receiver must re-synchronous their key generators before they can proceed further.

9.1.1 Classification of Stream Ciphers The symmetric stream encryption systems are classified in 1. Synchronous stream ciphers, and 2. Self-synchronizing stream ciphers. The sender can be found on the left and the receiver on the right side. When a synchronous stream cipher is used, the sender and the receiver of an encrypted message have to compute the keystream zt synchronously at any time t ≥ 0 for encryption and decryption 9.1.1.1 Synchronous stream ciphers Figure 3 depicts a symmetric, synchronous stream encryption system. The sender can be found on the left and the receiver on the right side. When a synchronous stream cipher is used; the sender and the receiver of an encrypted message have to compute the keystream z ciphers for time t, synchronously at any time t ≥ 0 for encryption and decryption. The keystream zt is generated independently from the plaintext message and the ciphertext. The encryption of the message symbols mt, t ≥ 0, can be described by the following equations: σt+1 =f(σt, k), zt =g(σt, k), ct =h(zt,mt),Where t≥ 0 is valid. The system has a state variable σt whose initial state σ0 can either be known publicly or be determined from the secret key k. In order to be able to carry out the encryption, the encryption function h must clearly be invertible. The function f is called the next state function and g is called the output function. The functions f, g and h are known publicly. 9.1.1.2 Self-synchronizing stream ciphers Besides synchronous stream ciphers, there are also self-synchronizing stream ciphers, but they are hardly used in information and communication systems. In this case the keystream zt depends on the key k and a fixed number l of previously generated ciphertext symbols.

154

Cryptographic Algorithms for Next Generation Wireless Networks Security

Figure 3. Synchronous, symmetric stream cipher

The encryption of a sequence of plaintext symbols mt, t ≥ 0, can be described by the following equations: σt = (ct−l, ct−l+1,..., ct−1), zt = g(σt, k), ct = h(zt,mt). Stream ciphers have several advantages which make them suitable for some applications. Most notably, they are usually faster and have a lower hardware complexity than block ciphers. They are also appropriate when buffering is limited, since the digits are individually encrypted and decrypted. Moreover, synchronous stream ciphers are not affected by error-propagation.

9.2 Block Cipher Encryption systems can be subdivided in symmetric and asymmetric systems as well as in block and stream ciphers. When block ciphers are used, a long message m is divided into blocks m = m0, m1, . . ., mN−1 of the same length. Here the blocks have usually a length of n = 64, 128 or 256 bits, depending on the processing length n of the block cipher. Padding mechanisms are used to fill the last block when the message m is not long enough so that the last block mN−1 is also an n-bit block. Then the single blocks mt, 0 ≤ t ≤ N –1 are assigned to a time-invariant encryption function f in order to obtain ciphertext ct = Ek(mt), where k is the secret, symmetric key. The parameters block length n and key length l should be chosen at least so large that a data complexity of 2n as well as a processing complexity of 2l is large enough not to allow an attacker to carry out an exhaustive key search in 10 or 20 years. Today, a block length of n = 64, 128 and 256 bits and an equally sized key length are used. The Feistel cipher is based on the idea of using the same function

155

Cryptographic Algorithms for Next Generation Wireless Networks Security

G: GF(2)l× GF(2)n/2→ GF(2)n/2 for encryption as well as for decryption. The function G, for example, consists of a product cipher. Here we assume that n is even and l is the length of the key k or a sub-key derived from it. The plaintext block m of length n bit is split into two equally sized blocks L and R, each having a length of n/2 bit: m = (L,R).Thenthe ciphertext block c is, as shown in Fig. 4.2, put together from the block R andthe bitwise XOR operation of block L with the function value G(k,R): c = (R, L + G(k,R)) = (R,X) . Block ciphers are somewhat faster than stream cipher each time ‘n’ characters executed, transmission errors in one cipher text block have no effect on other blocks. Block ciphers can be easier to implement in software, because the often avoid time-consuming bit manipulations and they operate on data in computer-sized blocks. In the real world block ciphers seem to be more general (i.e. they can be used in any of the four modes, the modes are ECB, CBC, OFB, CFB).They have different structure like Feistel Network, for e.g. Kasumi and Clefia. In a Feistel cipher (see Figure 4), the plaintext is split into two halves. The round function is applied to one half, and the output of the round function is bitwise ex-ored with the other half finally, the two halves are swapped, and become the two halves of the next round. Another is Substitution-permutation network e.g. AES. In Substitution-Permutation (SPN) cipher, the round function is applied to the whole block, and its output becomes the input of the next round. Although both stream ciphers and block ciphers belong to the family of symmetric encryption ciphers, there are some key differences. Block ciphers encrypt fixed length blocks of bits, while stream ciphers combine plain-text bits with a pseudorandom cipher bits stream using XOR operation. Even though block ciphers use the same transformation, stream ciphers use varying transformations based on the state of the engine. Stream ciphers usually execute faster than block ciphers. In terms of hardware complexity, stream ciphers are relatively less complex. Stream ciphers are the typical preference over block ciphers when the plain-text is available in varying quantities (for e.g. a secure wifi connection), Figure 4. Encryption principle of a Feistel cipher with one round

156

Cryptographic Algorithms for Next Generation Wireless Networks Security

because block ciphers cannot operate directly on blocks shorter than the block size. But sometimes, the difference between stream ciphers and block ciphers is not very clear. The reason is that, when using certain modes of operation, a block cipher can be used to act as a stream cipher by allowing it to encrypt the smallest unit of data available.

10. CONCLUSION AND FUTURE WORK A cryptographic algorithm is an essential part in network security. Most of the sensitive information in the wireless communication has latent security problems. An End-to-end security has been an issue in next generation wireless networks and hence a solution has to be proposed for the same using Secure Socket Layer/ Transport Layer Security (SSL/TLS), Virtual Private Network (VPN), or a similar mechanism should be provided for security of data. Cryptographic algorithms are utilized for security services in various environments in which low cost and low power consumption are key requirements. Wireless Local Area Networks (WLAN), Wireless Personal Area Networks (WPAN), Wireless Sensor Networks (WSN), and smart cards are examples of such technologies. Security is the most important part in Next Generation Wireless communication system, where more randomization in secret keys increases the security as well as complexity of the cryptography algorithms.

REFERENCES Al Tamimi. (2006). Security in Wireless Data Networks: A Survey Paper. Academic Press. Anjum, N., & Salil, K. (2009). Article. Authentication and Confidentiality in Wireless Ad Hoc Networks, 21, 28. Ayu, T., & Sudip. (2007). A multi-factor security protocol for wireless payment- secure web authentication using mobile devices. IADIS International Conference Applied Computing. Chandra, D.V., Shekar, V.V., Jayarama, & Babu. (2008). Wireless security: A comparative analysis for the next generation networks. Journal of Theoretical and Applied Information Technology. Como, L. H., Cortier, V., & Zalinescu, E. (2009). Deciding security properties for cryptographic protocols. Application to key cycles. ACM Transactions on Computational Logic, 5, 1–38. Daemean, J., & Rijmen, R. (1999). AES Proposal: Rijndeal version 2. Available at http://www.esat. kuleuveb.ac.be/rijmen/rijndeal Feng. (2008). Attacking Antivirus. Nevis Networks,Inc. Heys, H. M. (2001). A Tutorial on Linear and Differential Cryptanalysis. Academic Press. Kishor, S., Trivedi, Y. Z., Ma, & Dharmaraja, S. (2003). Performability modeling of wireless communication systems. International Journal of Communication Systems, 16, 561–577.

157

Cryptographic Algorithms for Next Generation Wireless Networks Security

Kumar, Y., & Prashant. (2010). Hardware Implementation of Advanced Encryption Standard. In Proceedings of International Conference on Computational Intelligence and Communication Networks. IEEE. 10.1109/CICN.2010.89 Liang & Wang. (2004). On performance analysis of challenge/responsebased authentication in wireless networks. Elsevier. Mohammad, S. (2013). A New Secure Cryptography Algorithm Based on Symmetric Key Encryption. Journal of Basic and Applied Scientific Research. Retrieved from www.textroad.com Muhammad & Seon. (2007). Security Issues in Wireless Mesh Networks. In Proceedings of International Conference on Multimedia and Ubiquitous Engineering (MUE’07). IEEE. Okumoku-Evroro &Oniovosa. (2005). Internet Security: The Role Of Firewall System. Department Of Computer Science Delta State Polytechnic Otefe-Oghara. Preneel, B. (2003). Analysis and Design of Cryptographic Hash Functions. Academic Press. Protecting the confidentiality of Personal Data. (2008). CMOD Department of Finance. SANS Institute (2005) “SANS Institute Info Sec Reading Room” Singh & Jain. (2012). Research Issues in Wireless Networks. International Journal of Advanced Research in Computer Science and Software Engineering, 2(4). Sun & Liu. (2004). Scalable Hierarchical Access Control in Secure Group Communications. IEEE. Taiming, F., Chuang, W., Wensheng, Z., & Lu, R. (2001). Confidentiality Protection for Distributed Sensor Data Aggregation. Retrieved from http://www.cs.iastate.edu/ Vinay. (2007). Authentication and Access Control The Cornerstone of Information Security. Trianz White Paper. Webroot Software Inc. (2004). Anti-spyware software: Securing the corporate network. Academic Press. Yagu¨e, Mana, & Lopez. (2005). A metadata-based access control model for web services. Retrieved from www.emeraldinsight.com/researchregister

This research was previously published in Next Generation Wireless Network Security and Privacy edited by Kamaljit I. Lakhtaria; pages 265-285, copyright year 2015 by Information Science Reference (an imprint of IGI Global).

158

159

Chapter 10

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network Alka Prasad Sawlikar RCERT Chandrapur, India Zafar Jawed Khan RCERT Chandrapur, India Sudhir Gangadharrao Akojwar Government College of Engineering, India

ABSTRACT To reduce communication costs, to protect our data from eavesdropping and from unauthorized users, cryptographic algorithms are used. Cryptographic module has to be developed for combining the operation of compression and encryption synchronously on the file. The information file is preliminary processed and then converts into one intermediary form so that it can be compressed with better efficiency and security. In this paper an optimized approaching coding technique which deals with both the issues of size and security is introduced and characterized experimentally using the performance measurement approach java in which file of any data length can be practically compressed and encrypted using new encryption technique and a novel energy saving technique in wireless communication network with efficient hardware solution is presented. To improve the strength and capability of algorithms and to compress the transmitted data an intelligent and reversible conversion technique is applied.

DOI: 10.4018/978-1-7998-1763-5.ch010

Copyright © 2020, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

INTRODUCTION From last consecutive years we have seen an unrivalled explosion in the amount of information or text data which is transmitted via many digital devices and for reducing the traffic, there is a need of strong cryptographic techniques so that large amount of information can be transmitted. For this a number of novel compression algorithms have been proposed earlier such as LZW, RLE, DWT, DCT and HUFFMAN. However, few of the above algorithms have been able to achieve best compression ratio. Advantage of security is to ensure that our information remains confidential and only access by authorized user and ensure that no one has been able to change that one, so it provide full accuracy. Compression is used to compress and secure the data, because it uses less space and saves money. It increases speed of data transfer from disk to memory. Requirements for data security are confidentiality, authentication, integrity and freshness. It involves transforming data of a given frame, called source message to data of a reduced sized frame called code word. There are different security techniques which are existing like AES, DES, ECC, RSA. Cryptography is valuable for protecting sensitive data online, especially in a world in which an increasing more systems are connected and unsafe to outside attack. It is also a valuable tool for authentication, allowing a user to verify his identity and statements using a public key encryption system. The main advantage of cryptography is as a security tool. Because any system connected wirelessly is bound to eventually be attacked by adversary, and it can be extremely tough to create a system that is invulnerable to outsiders. However, the mathematical analysis involved in encryption is complex enough that even if enemies manage to steal an encrypted file; he may never be able to break the code and access the contents. Strong encryption can be a last line of defense against outsiders, and can protect data even when it is being transferred through a connection that is not secure. The public and private keys associated with public key cryptography which offers unique advantages to their users like if a user encrypts data with his private key, anyone can get original intelligent data with his public key, verifying that he and only he could have sent the transmission. A public key can also encode data that only that specific user can decode, creating secure one-way communications on the Internet. Currently compression and encryption methods are doing simultaneously. Combination of two processes into one provides more security by this hybridization. While combining both compression and encryption data will be first compressed using compression techniques and then encryption techniques will applied and then comparative analysis will be done. If encryption and compression are done at the same time then it takes less processing time and more speed. Security of information is always been in demand since past few years and plenty of occurrences highlight the importance of the security of text data. As it is known, cryptography is a skill of hiding data and has been known from a long time, e.g. credit cards, debit cards, saving accounts, important documents and what not, everything needs protection. The most principal issue in world today is the large amount of valuable information that is flowing among various networks and present network development demands swap of information with more compression and security in both the time and space for data transmission for data storage (Jain, Lakhtaria, & Srivastav, 2013). This can be done by compression and encryption, such type of scheme is known as encryption compression crypto scheme. This ciphering or encryption is indeed a secure coding technique, whose purpose is to reduce the space for data storage and

160

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

time for data transmission and thus compression ratio becomes an most important parameter which we have to always keep in mind. So the data, which needs to be protected, is increasing on a fast and rapid rate and can be handled a bit if we can reduce its size or can remove the redundancy and for this both encryption algorithm and compression technique have to be merged, combined and made them work on precious information so that our message or file will be of compressed and encrypted form and secure and is easy to handle because of its reduced size and encrypted form which extends many advantages like saves space, manageable, easily transferrable, practical, and feasible. Take a look around we observed that the rapid adoption of network technologies and computing systems has brought newer risks and threats such as stealing, unauthorized access service, interruptions and reversing of information, and so on. This becomes more intense to suggest the importance of security and presents every organization with the legal and ethical responsibility to properly secure its information by using appropriate processes and measures. Imposing security at all states promises that information is processed, stored, or transmitted with integrity, reliability, and authenticity and available to all authorized entities. For a protected and reduced transfer of text information, the algorithms of compression must be combined with the encryption of text data. The techniques of compression removes the redundancies contained in the text file in order to reduce the amount of information and on the other hand, the techniques of encryption Due to its precious facts like object oriented, platform independent, simple, secure, portable, robust, architectural neural, multi-threaded, interpreted, dynamic, high performance, distributed etc JAVA is highly used for developing software. Here we are using Eclipse editor developed by eclipse open source community. (Dey, 2017) Some applications required to transmit same message or information which contains signal over channels. When there is a high percentage of unreliability of transmission of any information then at that time some special types of modes of transmission should be selected. So, to overcome the effects of collision or interference multichannel signaling is introduced in wireless communication systems. By doing so one can transmit same message or information which contains signal over channels so that the information can be recovered. To reduce storage cost by eliminating redundancies which occurs in number of files, data compression techniques are used. For storage of data and for transmission we have to pay money and this cost increases with the quantity of data available. But this cost can be reduced by processing the data so that it takes less transmission time and less memory. Well-known data compression algorithms are available and this paper deals with various data compression and encryption algorithms. The analysis of these algorithms can be used for finding various parameters which is applied in this proposed system. The new data compression technique has been proposed in this paper which is based on bit quantization level. MATLAB simulation for the best crypto system is shown and the results were observed. This crypto system is then applied to wireless communication network using NS2 software. Observation shows that the new proposed compression technique when combined with cryptographic technique gives the best results in terms of various parameters like delay, throughput and energy. New wireless sensor network (WSN) applications are growing up with recent technological advances in wireless communication and micro-electronics. The emergence of low cost multimedia devices and low-power such as microphones and micro-cameras has stimulated the development of the next generation of WSNs, known as wireless multimedia sensor network (WMSN).

161

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

The use of WMSN suggests a big interest for a wide variety of applications such as object detection, localization, tracking and counting. In fact, these systems can be applied to assist the elderly people, to intensify and complement existing surveillance systems against crime and to extend the ability of law enforcement agencies to monitor areas, public events, private properties and borders (Kaddachi & Soudani, (2010)) The data transmission, as a potential application on WMSN, is still limited to some problems related to energy consumption and storage capacity in nodes, as well as to the available bandwidth of the wireless links. Nodes in WMSN should ensure ad-hoc routing packets for other communication processes, so when sensing image or data, the CPU’s of [INSERT FIGURE 001]these nodes can be heavily loaded with data processing tasks, and may be out of their capacities (Culurciello & Andreou, 2006; Akyildiz, Tommaso, & Chowdhury, 2007) The data compression at the source node is being considered to be the efficient solution to protect the physical capacities of the node and to enhance its life-time. However, a big attention should be paid when building up an efficient and strong compression scheme that avoid weighty processing tasks falling down the available energy and making the node out of its capacities to relay with its neighbors. A big mismatch comes into sight between the complexities of these compression algorithms and the available physical resources of WSN node for TinyOS-based implementation. However, most recent sensor platforms are software-based, as is the Cyclops platform. Some works have shown that popular algorithms such as JPEG, JPEG2000 or SPIHT are generally not efficient in software simulations because they lead to greater energy consumption and storage than the transmission of the uncompressed image. That is due to the stock limitation of the software-based platforms in terms of processor speed. A hardware solution for data compression at source node can achieve dynamic compression rate according to the end user requirements. This solution represents a hardware circuit, which is intended to be embedded in the source and destination node. It will be considered as a co-processor for tasks related with compression to unload the node microcontroller giving it more computational bandwidth for wireless network interactions. Further, it will enhance time processing and achieve low power consumption. (Akyildiz, Tommaso, & Chowdhury, 2007; Duran-Faundez & Lecuire, 2008)

LITERATURE SURVEY Several researchers have been carried cryptography. In literature survey found various compression and encryption algorithms for finding different energy efficient parameters. But still not found satisfactory solution for said problem so developed an algorithm for efficient energy saving. Following are some cryptographic techniques which have been studied. In the paper “A Comparative Study of Lossless Compression Algorithm on Text Data,” Jain, Lakhtaria, & Srivastav (2013) shows the comparison of different lossless compression algorithm over text data which contain different text patterns and had drawn compression ratio of all the algorithm by considering the compression time, decompression time and from comparison they concluded that the Huffman Encoding is considered as the most efficient algorithm. Authors have taken different file sizes using RLE, Adaptive Huffman and Shannon Fano algorithms done comparison based on compression ratio and found how compression ratio getting varied and observed that the size of compressed file created by Adaptive Huffman algorithm is very less in comparison to other algorithms (Jain, Lakhtaria, & Srivastav, 2013).

162

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

Med Lassaad Kaddachi, Adel Soudani (2010), in the paper “Efficient hardware solution for low power and adaptive image-compression in WSN”, have presented a hardware solution for low power image compression which is based on Loeffler DCT transform. This solution is intended to be machine embedded as a coprocessor to the main WSN CPU. This solution provides powerful compression settings at run time according to the node state and the user requirements. It saves energy and grants short processing time compared to a software implementation of the image compression scheme. Authors have concluded that this solution can be enhanced with the integration of a CMOS image sensor that will reduce the energy and the acquisition time of the image. With this the image is compressed and then avoids energy loss and lengthy processing time. They proposed architecture for compression circuit at the source node and outline the image quality analysis and provide the characteristics of the proposed solution using FPGA and CMOS Standard-Cell and finally done comparison with the TinyOS by implementing. (Kaddachi & Soudani, 2010) E. Culurciello, and A.G. Andreou, in “CMOS image sensors for sensor networks,” discusses about two generations of CMOS image sensors with digital output fabricated. Collision on the output is displayed using contention detector circuits. The image sensors present very high dynamic range and ultra-low power operation. These characteristics allow the sensor to operate in different illuminating conditions and for years on the sensor network node power budget. Authors have concentrated on power saving technique of sensor nodes. (Culurciello & Andreou, 2006.) In “A survey on wireless multimedia sensor networks,” Ian F. Akyildiz, Tommaso Melodia, Kaushik R. Chowdhury, outlined the main research challenges. Algorithms, protocols, and hardware for the development of WMSNs were surveyed, and open research issues discussed in detail. Authors classified currently off-the-shelf hardware as well as available research prototypes for WMSNs. Furthermore, they discussed current solutions and open research issues at the application, transport, network, link, and physical layers of the communication stack, along with possible cross-layer synergies and optimizations. Further pointed out how recent work undertaken in Wyner–Ziv coding at the application layer, specialized spatio-temporal transport layer solutions, delay bounded routing, multi-channel MAC protocols, and UWB technology and amongst others, seems most promising research directions in developing practical WMSNs. (Akyildiz, Tommaso, & Chowdhury, 2007) N. Dey, A.S. Ashour, in book “Intelligent Techniques in Signal Processing for Multimedia Security”, proposes new algorithms to ensure secured communications and prevent unauthorized data exchange in secured multimedia systems. Focusing on numerous applications algorithms and scenarios, it offers an in-depth analysis of data hiding technologies including watermarking, cryptography, encryption, copy control, and authentication. The authors present a framework for visual data hiding technologies that resolves emerging problems of modern multimedia applications in several contexts including the medical, healthcare, education, and wireless. (Dey, 2017) Duran-Faundez, and V. Lecuire, in “Error Resilient Image Communication with Chaotic Pixel Interleaving for Wireless Camera Sensors,” discusses about new applications of wireless sensor networks which requires vision capabilities. They had considered the high loss rates found in sensor networks and the limited hardware resources of current sensor nodes, low-complexity robust image transmission must be implemented, avoiding as much as possible the need for retransmission or redundancy. They proposed a pixel interleaving scheme based in Torus Automorphisms, thus, neighboring pixels are transmitted in different packets. Hence, if packets are lost, authors found a high probability of retrieving enough information to obtain an approximation of the original value. Their results showed an increase of the image quality in comparison with a sequential raw image transmission approach, while preserving similar energy consumptions, time and low-complexity. (Duran-Faundez & Lecuire, 2008.) 163

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

In “Data compression algorithms for energy-constrained devices in delay tolerant networks,” Sadler, Christopher M. and Margaret Martonosi, focused on collecting additional significant energy improvements by planning computationally-efficient lossless compression algorithms on the source node which reduces the amount of information that must be passed through the network to the sink, and thus have energy advantages that are multiplicative with the number of hops the data travels through the network. Upon observation, they comment if sensor system designers want to compress obtained data, they must either develop application-specific compression algorithms or use off-the-shelf algorithms not designed for resource-constrained sensor nodes. In this paper, they discusses the design issues involved with implementing, adapting, and customizing compression algorithms specifically made ready for sensor nodes. While enlarging sensor LZW (S-LZW) and some simple, but effective, variations to this algorithm, they showed how different amounts of compression can lead to energy savings on both the compressing node and throughout the network and that the savings depends heavily on the radio hardware .So to validate and evaluate their work, they had applied it to datasets from several different real-world deployments and showed that their approaches can reduce energy consumption by up to a factor of 4.5X across the network.(Sadler & Martonosi, 2006.) Made Agus Dwi Suarjaya (2012) in “A New Algorithm for Data Compression Optimization”, presented that when the storage nears it limit, they then try to reduce those files size to minimum by using data compression software so they proposed a new algorithm for data compression, called jbit encoding (JBE) which manipulates each bit of data inside file to minimize the size without losing any data after coding which is classified to lossless compression. This basic algorithm is intended to be combining with other data compression algorithms to optimize the compression ratio. The performance of their algorithm is measured by comparing combination of different data compression algorithms. (Suarjaya, 2012) After detail study of above literature, design different algorithms for finding better solutions.

PROPOSED DATA COMPRESSION AND ENCRYPTION ALGORITHM This section presents a new compression and encryption technique which was not proposed earlier. Following is the proposed algorithm which compresses and encrypts the message: Step 1: Create a table for encoding the input symbols. The upper bound limit and lower bound limit of each new symbol can be calculated or values can assigned as a. Initially load lower bound limit = 0, upper bound limit = 1 b. Code all the symbols like a, b, c, d….. c. Find new values for current range, upper limit and lower limit Current range = upper bound limit – lower bound limit higher_ limit = lesser limit + (current range * higher limit of new symbol) lesser_ limit = lesser limit + (current range * lesser limit of new symbol) Step 2: The string can be encoded by fetching any value within the range of probability and after that convert the output decimal number into its binary format. Step 3: Limit the number of bits by using the formula and anyways cancel it or floor it or put it in bracket 164

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

Table 1. Representation of characters along with its probability occurrence Symbol

Probability

Range (lower limit, upper limit)

a

40%

(0.00, 0.40)

l

15%

(0.40, 0.55)

k

25%

(0.55, 0.80)

p

20%

(0.80, 1.00)

Also, No of bits = log [2/upper bound limit of last encoded symbol - lower bound limit of last encoded symbol] Step 4: In this way compression can be finished and the number of bits are used to reduce the number of bits obtained in step2. Step 5: Now for encryption choose any one binary key pad and EX-OR it with above. Step 6: Then rotate right 4 bits means we will get rotating resulted bits. Step 7: Convert above result into decimal format again. The result which we will get output is floating point number and that is corresponding to the input symbol. Now here is the algorithm which decompresses and decrypts the text Step 1: Convert the received data and information into binary form. Step 2: Rotate bits to left. Step 3: Selected binary key pad and EX-OR it with above result. Step 4: Convert the result back into decimal form and result should be noted as encoded value. Step 5: Encoded_value or Coded value = Encoded input Still string is not fully decoded so match the symbol containing encoded value within its range Current range = higher limit of new symbol - lesser limit of new symbol Encoded value = (encoded value - lower bound limit of new symbol) ÷ current range at the output we will found the original string.

DETAIL MATHEMATICAL ANALYSIS OF AN ALGORITHM Take following example to explain the above algorithm in a better way and following Table 1 represents number of characters with its probability. Compression and Encryption have been done in a following manner. Data to be encoded and encrypted is “alkp” Step 1: Code ‘a’ as current_ range = 1 - 0 = 1 upper limit bound = 0 + (1 × 0.4) = 0.4

165

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

lower limit bound = 0 + (1 × 0.0) = 0.0 Code ‘l’ current_ range = 0.4 - 0.0 = 0.4 upper limit bound = 0.0 + (0.4 × 0.55) = 0.22 lower limit bound = 0.0 + (0.4 × 0.4) = 0.16 Encode ‘k’ Current_ range = 0.22-0.16 = 0.06 upper limit bound = 0.16 + (0.06 × 0.8) = 0.208 lower limit bound = 0.16+ (0.06 × 0.55) = 0.193 Encode ‘p’ current _range = 0.208-0.193 = 0.015 upper limit bound = 0.193 + (0.015 × 0.90) = 0.2065 lower limit bound = 0.193+ (0.015 × 0.80) = 0.205 Step 2: The string “alkp” may be encoded by any value within the range [0.2065, 0.205]. Now our output is 0.20425 whose binary equivalent = 0.0011010010010011011101 Step 3: No_ of_ bits = [log_2/0.0015] = [log1333.33] = 8bits Step 4: So, after minimizing number of bits binary value is 0.00110100. Step 5: Let Our One time key pad is – 0.10101010 Data- 0.00110100 from step 4. After EX-ORing we get the output as 10011110 Step 6: Rotate 4 bits right the result is 11101001 Step 7: 0 .11101001 in decimal is 0.91015625 Decompression and Decryption have been done in a following manner: Step 1: Received main data is 0.91015625 and binary format of received data is 0.11101001 Step 2: Apply 4 times left shifts to result of step1 the result is 10011110 Step 3: Apply selected one time key pad i.e. 10101010 and logical EX-OR it with the result of step2 the result is 0.00110100 Step 4: Convert 0.00110100into decimal i.e. 0.203125 Step 5: See the probability ranges from table and decodes the four-character string encoded as 0.203125 Decode first symbol ‘a’ 0.20312 5 is within [0.00, 0.40) 0.203125 encodes ‘a’ Remove effects /changes of ‘a’ from encode value Current _range = 0.40 - 0.00 = 0.40 Coded or Encoded _value = (0.203125- 0.0) ÷ 0.40 = 0.50775 Decode next symbol ‘l’ 0.50775is within [0.40, 0.55) 0.50775 encodes ‘l’ Remove effects/changes of ‘l’ from encode value current range = 0.55 - 0.40 = 0.15 encoded value = (0.50775 - 0.40) ÷ 0.15 = 0.71833 Decode next that is third symbol ’k’ 0.71833is within [0.55, 0.80) 0.71833encodes’k’ 166

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

Remove effects/changes of ‘k’ from encode value current range = 0.80 - 0.55 = 0.25 encoded value = (0.71833-0.55) ÷ 0.25=0.67332 Decode next third symbol ’p’ 0.67332 is within [0.80, 0.90] 0.67332 encodes ‘p’

IMPLEMENTATION OF NEW SYMMETRIC CRYPTOGRAPHIC TECHNIQUE USING JAVA Computer security is concerned with the preservation of information in environments where there is a possibility of intrusion or malicious action. A vital technology that underpins the security of information in computer networks is known as Cryptography. The new trend in cryptography emphasizes the design of efficient cryptographic primitives that are provably secure in the standard model. For improving security systems government agencies are investing a considerable amount of resources because of recent terrorist activities that dangerously exposed flaws and weaknesses in today’s system safety mechanisms. Java is an object-oriented programming language with its own runtime environment. This is originally developed by Sun Microsystems and now merges into ORACLE Corporation. They can be run on any operating system that’s why said as platform independent. Java code is famous as WORA that is writing once run anywhere. Now a day’s three billion devices are running java as it is adopted as platform independent programming language. It is used to develop dynamic web pages, in standalone application, enterprise application as well as in mobile. (Bisht & Singh, 2015) Due to its precious facts like object oriented, platform independent, simple, secure, portable, robust, architectural neural, multi-threaded, interpreted, dynamic, high performance, distributed etc it is highly used for developing software. Here we are using Eclipse editor developed by eclipse open source community. (Idrizi, Florim, Dalipi et al., 2013)

STEPS OF SYMMETRIC CRYPTOGRAPHIC ALGORITHM Encryption Algorithm 1. Take input in the form of text. This message can be any text file with random data size including characters, symbols and numbers. 2. Operations are performed on every symbol of input: Stream cipher. 3. Generate any key of any length. This key should be same for both encryption and decryption. 4. Perform XOR operation of message with key but not directly. First calculate length of that key used which is constant throughout and take modulus of key length and each key character. Note the remainder. 5. Then XOR remainder with each character of message. Thus, the message will get encrypted.

167

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

Decryption Algorithm Only single secret key has been exercised in the encryption. It is a symmetric algorithm, so same key is used for encryption and decryption. Decryption method of the algorithm is the reverse of their encryption method. 1. Take input as encrypted message. 2. Key is same 3. Perform XOR operation of message with key but not directly. First calculate length of that key used which is constant throughout and take modulus of key length and each key character. Note the remainder. 4. Then XOR remainder with each character of message. Thus, message will get decrypted.

Stages for Program The following are the essential steps for reading file, compressing and encrypting Public void start compression (String file Path){ //Read File Array List data Into ASCII = read File(file Path); //Compress File String compressed Data = compression Algorithm(data Into ASCII, file Path //Encrypt File String encryptedData = EncryptDecrypt.getInstance(). encryptData(compressedData); //Write final encrypted file WriteDataIntoFile(encryptedData,filePath,”FinalEncryptedData.txt”);}

IMPLEMENTATION OF PROPOSED ENCRYPTION ALGORITHM The above are the essential steps for reading file, compressing and encrypting. In encryption and decryption process we had taken single key and XOR operation. The reason behind it is if we used multiple keys and more operations then time as well as size of encryption will increase. Read file is a method to go to declaration where the logic is written which will convert each character into corresponding ASCII number. When it enters into declaration for read file it will check every character in the line and also checks whether that file is present in that path and if not then it will give error. Thus, the file reading process is going on and once it will finish we can observe size as above and then return the ASCII values. This ASCII data have to send to compression algorithm along with file path and finally this method will return compressed data. Compression algorithm will fetch ASCII along with file, logically compressed and add into compressed file and send for encryption.

168

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

Since encryption, decryption and decompression is created in another file, we have to call the algorithm only. Then in encryption process first set key, give null, means void means nothing no compressed message neither key is present so no performance will occur so it has to be confirmed that both key and message should be available. Once it is confirm then split it into character, observe length in integer and then XOR. Save as base 64 encode which is a file where actual encoding takes place and then performs encryption on it. For decryption reverse process occurs, decrypted data send for decompression we will get decompressed ASCII data then finally convert ASCII integer to original data. Decryption is done on direct encrypted file only with same key which is used in encryption then send to base 64 decoder then XOR so that we will get exactly decoded data which we had used for encrypted. For implementation any file of infinite length can be chosen. Using compression and encryption technique we get following results. Figure 1 shows text file fetched from notepad, Figure 2, Figure 3, Figure 4, Figure 5 shows fetched file with size, ASCII, compressed and encrypted values. With the help of programming we can see its size, ASCII values, compressed file, encrypted data and decrypted data as follows: Similarly, we can observe size of every decrypted file, decompressed data and ASCII to original file in reverse manner as follows. Figure 6, Figure 7, Figure 8 and Figure 9 shows the values of decrypted file, decompressed data and ASCII to original.

INNOVATIVE ENERGY SAVING TECHNIQUE IN WIRELESS NETWORK To reduce storage cost by eliminating redundancies which occurs in number of files, data compression techniques are used. As we know two types of data compression techniques are there: one is lossy and another is lossless and for reducing file size after decoding, lossy compression technique is used for video, audio and text compression. For storage of data and for transmission we have to pay money and this cost increases with the quantity of data available. But this cost can be reduced by processing the data so that it takes less transmission time and less memory space. Different data types consist of many lumps of repeated data. Such fresh data can be altered into a compressed data which saves a lot of storage and transmission costs. Figure 1. Text File fetched from Notepad

169

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

Figure 2. Fetched file with size, ASCII, Compressed and Encrypted values

Figure 3. ASCII values of original file

170

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

Figure 4. Compressed data of original file

Figure 5. Encrypted file after compression using proposed encryption technique

Sensor network establishes sensor network between neighboring nodes with radio range transmission and during communication one node exchanges their information to the neighboring node and thus easily discovers new nodes. In addition, a power source often consists of a battery with a limited energy budget. In a well-designed network, the sensors in a certain area exhibit homogeneous behaviors to achieve energy balance. In other words, when one sensor dies, it can be expected that the neighbor of this node will give out energy very soon, since they will have to take the responsibilities of that died sensor and expects the lifetime of several months to be converted into several years. Thus, energy saving is crucial in designing life time wireless sensor networks.

171

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

Figure 6. Values of decrypted file, decompressed data and ASCII to original

Figure 7. Decrypted data file with new symmetric cryptographic algorithm

172

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

Figure 8. ASCII values of decompressed data

Figure 9. ASCII to Original converted file

Advanced Compression and Decompression Algorithm The different steps of the proposed algorithm are: 1. Take input data string of random length or specified length. 2. Decimate the data length by 2 3. Convert decimal data string to binary 4. Find the polarity of consecutive input sample as If next_sample > current_sample Make LSB bit of current output as ‘1’

173

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

Else make LSB bit of current output as ‘0’ 5. Convert binary data string to decimal 6. We will get compressed output data string For decompression: 1. 2. 3. 4.

Take first compressed output bit as it is. Consider two consecutive bits and check current_sample and next_sample Convert decimal data string to binary If current_sample has 1 in LSB position then current_sample is always equal to new_output else add two consecutive samples and divide it by 2 that will be new_sample 5. Convert binary data string to decimal. Repeat the same process for complete data length; we will get decompressed output bits. The length of input bits will always equal to length of decompressed data bits and this algorithm is named as KSA compression algorithm.

Experimental Analysis and Results Energy optimization is carried out over channels using security protocol SSL(Socket Secure Layer) and NS2 software. The number of nodes can be varied, start and stop time can also be varied and can observe various parameters like energy, delay and throughput. Table 2 shows energy comparison table when protocol is not applied and using KSA for nodes=30. Figure 10, Figure 11 and Figure 12 shows the graphical representation of comparative energy graph, delay graph and throughput graph. Table 2 shows if we go on increasing simulation time at one instant as compare to standard energy parameters KSA is giving less values and hence is energy efficient.

Table 2. Energy Comparison table when protocol is not applied and using KSA for nodes=30 Simulation Time

O-Energy

Energy

KSA_Energy

0

110

82

50

5

200

142

92

10

210

170

99

15

220

178

99

20

230

180

100

25

220

182

100

30

210

180

100

174

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

Figure 10. Comparative energy graph with number of nodes = 50 and stop time = 0.1sec

Figure 11. Comparative delay graph with number of nodes = 50 and stop time = 0.1sec

175

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

Figure 12. Comparative throughput graph with number of nodes = 50 and stop time = 0.1sec

Limitations and Future Scope Lossy compression was implemented in the complete work where editable data was dumped, difficult to edit and while transmitting through protocols it was taken large amount delay. So we could design new lossless compression algorithm which when combines with encryption techniques the values of the parameters will get increased and also while transmitting over wireless communication network we can use new transmission technique, send data using protocol and can find parameters like delay, energy and throughput.

CONCLUSION A safe and secure application of compressed cryptographic technology will pay attention to how hackers are detected keys and associated with user identities. The compression algorithm which has been introduced offers compression with less compression time. To strengthen the security of the communication network, this technique has been suggested and with the experimental results it has been proved that it leads to increase not only security but is also energy efficient. The complete suggested design makes the cryptanalysis difficult for the intruder which is achieved effectively through a module which is a perfect blend of new compression technique with cryptography principles. This design is really instrumental in

176

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

providing very big challenge to the intruders who attempt to break the algorithms by any means. The encryption technique introduced in this paper is implemented using Java by giving example and is showing practical results of how data gets read, compressed, converted in to ASCII, encrypted, what is exact size of original file, compressed, encrypted, decrypted and decompressed file. The complete suggested encryption algorithm makes the cryptanalysis difficult for the intruder which is achieved effectively through a module and is a perfect blend of new symmetric encryption technique with modular cryptographic principles. This design is really instrumental in providing very big challenge to the intruders who attempt to break the algorithms by any means. Thus, it can be concluded that the secret data can be transmitted securely in an insecure media using this module. Proposed algorithm is viewed best in terms of speed, cost, throughput, security and power consumption.

REFERENCES Akyildiz, I. F., Melodia, T., & Chowdhury, K. R. (2007). A survey on wireless multimedia sensor networks. Elsevier. Computer Networks, 51(4), 921–960. doi:10.1016/j.comnet.2006.10.002 Idrizi, F., Dalipi, F., & Rustemi, E. (2013). Analyzing the speed of combined cryptographic algorithms with secret and public key. International Journal of Engineering Research and Development, 8(2), 4. Bisht, N., & Singh, S. (2015). A Comparative Study of Some Symmetric and Asymmetric Key Cryptography Algorithm. International Journal of Innovative Research in Science. Engineering and Technology, 4(3), 102. Bisht, N., & Singh, S. (2015). A Comparative Study of Some Symmetric and Asymmetric Key Cryptography Algorithms. International Journal of Innovative Research in Science, 4(3), 10281. Brumley, B.B., & Hakala, R.M. (2009). Cache-timing template attacks. In ASIACRYPT (Vol. 5912, pp. 667-684). Celikel, E., & Dalkilic, M. E. (2004). Experiments on a secure compression. Proceedings of the International Conference on Information Technology: Coding and Computing ITCC ‘04 (Vol. 2, pp. 150-152). IEEE. Charfi, Y., Wakamiya, N., & Murata, M. (2009). Challenging issues in visual sensor networks. IEEE Wireless Communications, 16(2), 44–49. doi:10.1109/MWC.2009.4907559 Culurciello, E., & Andreou, A. G. (2006). CMOS image sensors for sensor networks. Analog Integrated Circuits and Signal Processing, 49(1), 39–51. doi:10.100710470-006-8737-x Dey, N. S. (2017). Intelligent Techniques in Signal Processing for Multimedia Security. Switzerland: Springer International. doi:10.1007/978-3-319-44790-2 Duran-Faundez., & Lecuire, V. (2008.). Error Resilient Image Communication with Chaotic Pixel Interleaving for Wireless Camera Sensors. Proceedings of the Workshop on Real-World Wireless Sensor Networks (REALWMSN’08), Glasgow, Scotland.

177

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

Jain, A., Lakhtaria, K. I., & Srivastav, P. (2013). A Comparative Study of Lossless Compression Algorithm on Text Data. Proceedings of the International Conference on Advances in Computer Science (p. 536). AETACS Elsevier Digital Library. Jain, Y. K., & Gosavi, P. B. (2008). Email Security using Encryption and Compression. Proceedings of the IEEE Int. Conf. of Comput. Intel. Model. Control Autom. (p. 136). Kaddachi, M. L., & Soudani, A. (2010). Efficient hardware solution for low power and adaptive imagecompression in WSN. Proceedings of ICECS (pp. 583–586). IEEE. Lenstra, A. K., Hughes, J. P., Augier, M., Bos, J. W., Kleinjung, C.T., & Wachter. (2012). Public keys. In Crypto, LNCS (Vol. 7417, pp. 626-642). Marcelloni, F. (2008). A Simple Algorithm for Data Compression in Wireless Sensor Networks. IEEE communications letters, 12(6), 411-413. Motgi, N., & Mukherjee, A. (2001). Network Conscious Text Compression Systems (NCTCSys). Proceedings of International Conference on Information and Theory: Coding and Computing, Las Vegas, USA. IEEE Computer Society. Nadeem, A. (2006). A performance comparison of data encryption algorithms. Proceedings of the First international conference on Information and communication technologies ICICT ‘05 (pp. 84-89). Patil, M. V. (2014). SMS text Compression and Encryption on Android O.S. Proceedings of the International Conference on Computer Communication and Informatics (ICCCI). 10.1109/ICCCI.2014.6921767 Sadler, C. M., & Martonosi, M. (2006.). Data compression algorithms for energy-constrained devices in delay tolerant networks. Proceedings of the ACM Conference on Embedded Networked Sensor Systems (SenSys). 10.1145/1182807.1182834 Sagheer, A.M., Al-Ani, M.S., & Mahdi, O.A. (2013). Ensure Security of Compressed Data Transmission. Proceedings of the Sixth International Conference on Developments in eSystems Engineering (pp. 270-275). IEEE Computer Society of India. 10.1109/DeSE.2013.55 Sangwan, N. (2013). Combining Huffman text compression with new double encryption algorithm. Proceedings of the IEEE International Conference on Emerging Trends in Communication, Control, Signal Processing & Computing Applications (C2SPCA). Savari, M., Montazerolzohour, M., & Thiam, Y. E. (2012). Combining Encryption Methods in Multipurpose Smart Card. Proceedings of the IEEE Int. Conf. CyberSec. Shahi, G., & Singh, C. (2013). Securing and Compressing Transmission over LAN by using Public Key Cryptography. International Journal of Innovative Research in Computer and Communication Engineering, 1(7), 2320–9798. Stallings & William. (2005). Cryptography and Network Security: Principles and Practices (4th ed., Vol. 8). Prentice Hall International .

178

Efficient Energy Saving Cryptographic Techniques with Software Solution in Wireless Network

Suarjaya, M. A. (2012). A New Algorithm for Data Compression Optimization. International Journal of Advanced Computer Science and Applications, 3(8), 14–17. Ukhopadhyay, B. A. (2010). Cryptography and network security (2nd ed.). New Delhi: Mc Graw Hill Education.

This research was previously published in the International Journal of Synthetic Emotions (IJSE), 7(2); edited by Amira S. Ashour and Nilanjan Dey; pages 78-96, copyright year 2016 by IGI Publishing (an imprint of IGI Global).

179

180

Chapter 11

Applicability of Cellular Automata in Cryptanalysis Harsh Bhasin Jawahar Lal Nehru University, India Naved Alam Jamia Hamdard, India

ABSTRACT Cryptanalysis refers to finding the plaintext from the given cipher text. The problem reduces to finding the correct key from a set of possible keys, which is basically a search problem. Many researchers have put in a lot of effort to accomplish this task. Most of the efforts used conventional techniques. However, soft computing techniques like Genetic Algorithms are generally good in optimized search, though the applicability of such techniques to cryptanalysis is still a contentious point. This work carries out an extensive literature review of the cryptanalysis techniques, finds the gaps there in, in order to put the proposed technique in the perspective. The work also finds the applicability of Cellular Automata in cryptanalysis. A new technique has been proposed and verified for texts of around 1000 words. Each text is encrypted 10 times and then decrypted using the proposed technique. The work has also been compared with that employing Genetic Algorithm. The experiments carried out prove the veracity of the technique and paves way of Cellular automata in cryptanalysis. The paper also discusses the future scope of the work.

1. INTRODUCTION One of the most important factors responsible for the development of human race is the ability to communicate. The development and the design of communication system has become one of the most contentious issues. The security of communication is, therefore, one of the most essential attributes in any communication system. The development in the field of cryptography has helped achieve the dream of a secured communication. However, the system becomes vulnerable in one of the following cases. The intruder might want to ‘listen’ to the communication, which is referred to as eavesdropping. The DOI: 10.4018/978-1-7998-1763-5.ch011

Copyright © 2020, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

Applicability of Cellular Automata in Cryptanalysis

intruder might want to change data or worse might use it for some other purpose. A sound cryptography technique prohibits any of these. One of the easiest methods of cryptography is to apply XOR function on the data (plaintext) and key. The resultant is referred to as cipher-text. The key when XOR-ed with the plaintext again produces plaintext (Bruce, 1995; & Rothe, 2002). The soundness of a system, though, is not easy to ascertain. The breakability of the key can be one of the major factors in determining the goodness of a system. One of the ways of doing so is to find how good a system is to diversified attacks. Given the ciphertext, the process of finding the plaintext is referred to as Cryptanalysis. Cryptanalysis has been one of the most researched topics in the field of Network Security. Various researchers have devised different methodologies to accomplish the task. Owing to the importance of the topic, it is therefore necessary to carry out an extensive literature review of the techniques and find the gaps therein. The review is also important to justify the applicability of soft computing techniques, especially Cellular Automata (CA), to handle the problem. The work intends to achieve the above goals. The goals of this paper are as follows. • • • •

To carry out a literature review of cryptanalysis using soft computing To find the gaps in the existing techniques To propose a technique using CA To verify and validate the technique

The paper has been organized as follows. The second section explains the literature review, the third section explains the concepts of CA, the fourth section explains the proposed work, the fifth section gives the results and the last section concludes. The work paves way of CA in cryptanalysis.

2. LITERATURE REVIEW Cryptography is one of the most researched topics in Computer Science. The topic is not only important in securing essential data from eavesdropping and theft but has also been used to win wars. The power of cryptography was demonstrated in World War II. Cryptanalysis is the crafting of key, given a set of data and corresponding encrypted code. The researchers developed many models for cryptanalysis during the Second World War. These models proved instrumental in proving a strategic edge to Britain. The development in the field continued there-after, when the world was divided into two groups, both wanting to gain as much information as possible from the other. The turn of events would remind the fraternity of the importance of the cryptography. The breaking of PURPLE by William Friedman, breaking of ENIGMA by Alan Turing, problems faced in accessing the contents of Bin Laden’s drive was the constant reminders of the importance of this topic. The conventional techniques of Cryptanalysis include frequency counts, in order to ascertain the most frequently used syllabi. The use of letters and words in English gives a cue of what to expect from a given text. Another technique of cryptanalysis uses the study and analysis of patterns. Some of the researchers have used side chain attacks for accomplishing the above task but the method works only in constrained environment.

181

Applicability of Cellular Automata in Cryptanalysis

The present work deals with the problem using the power of CA and Neural Networks. In order to put the things in proper perspective, a review of the techniques has been carried out. The review was done to find the gaps in the existing techniques. The summary of the review has been provided in Table 1. The review points to the fact that techniques like CA have seldom being used by researchers. Moreover, even those who used it have not performed requisite experimentation to verify and validate the work. Table 1. The summary of the review S. NO.

Name of Author

Techniques

1.

Bhateja, A

The method uses Genetic Algorithms to break Vigenere cipher. The technique uses Roulette wheel selection; two points cross over and cross mutation.

2.

Luthra J

The work uses the statistical data of the language as the premise. The work combines Genetic Algorithms with the Firefly Algorithm for cryptanalysis of the mono alphabetic substitution cipher.

3.

Shujun Li et al

The work breaks the permutation multimedia cipher. The work has been verified and validated using Quantitative cryptanalysis

4.

Omran, S.S.; Al-Khalid, A.S.; Al-Saady, D.M.

The work breaks then poly alphabetic substitution cipher (Vigenère cipher) by applying genetic algorithms. The work uses frequency analysis as its base

5.

Jun song et al.

The work assigns fitness values to the keys and produces higher fitness value keys. The work proves that a 4 round DES can be broken by the model.

6.

Uddin M. F. and Youness A. M.

The work uses Particle Swarm Optimization (PSO) in the cryptanalysis of classical simple substitution ciphers.

7.

Feng-Tse Lin and Cheng-yan kao

The work presents a cipher text only attack. Genetic Algorithm has been used to break the Vernam cipher. The proposed approach is a cipher-text only attack in which we don’t know any plaintext. The only thing we have to know is that the plaintext is an English document.

8.

Girish Mishra, Saravjeet Kaur

The work breaks transposition cipher using Hill Climbing, Simulated annealing and combination of these.

9.

Xingyuan Wang, Qian Wang, Yingqian Zhang

This technique uses a logistic map to generate the keys. The results of the paper points to the fact that this method is better as compared to differential attacks.

10.

Chengqing Li, Shujun-Li, Gonzalo alvarez, Guanrong chen, Kwok Tund Lo

The work uses circular bit shift and XOR operations to generate the data. The work uses Chaotic System for control of pseudorandom bit sequences

11.

Mohammed M. Alani

The work uses a Neural Network based model. The work has been verified for DES and triple DES.

12.

R. Manjula, R. Anitha

the work tests various algorithms and designed a new one. According to the author, the success rate of the is 70 to 75 percent.

13.

Maiya Din, Ashok K Bhateja, Ram Ratan

The work uses Geffe generator, which is a non-linear binary key sequence generator. In this technique, the initial states are the secret key bits that maintain the security of Geffe generator.

14.

S. Franciszek, et., al.

The paper proposes a symmetric key cryptographic design based on 1-D cellular automata with radius 1 and 2. The quality of the system has been tested using standard attacks.

15.

Jaecchul Sung, et., al.

This paper implements self-invertible cellular automata based structure for cryptanalysis and uses the conjugate property as the base.

16.

M. Tomassini, & Mathieu Perrenoud.

The paper proposes 1-D and 2-d non-uniform cellular automata base system for cryptanalysis and discuss its hardware implementation.

17.

Feng Bao

The paper combines affine transformation with cellular automata to produce cryptanalysis system whose security has been tested using standard attacks.

18.

H. Nhasin & N. Arora

The paper proposes genetic algorithm based key generation algorithm and explains the implementation of the system and discusses the results of various attacks on the system.

182

Applicability of Cellular Automata in Cryptanalysis

3. CELLULAR AUTOMATA The story of CA starts from “A New Kind of Science”, a book by Wolfram (1984). The book initiated an intestate in the field, which was reflected in the works of Ulam and Von Neumann (Ulam, 1970) and numerous research papers published thereafter. Since then, CA has been successfully applied to various fields ranging from Artificial Life to Sociology. This paper explores the applicability of this fascinating field in cryptanalysis. On the face of it, the concept appears naive to some and ‘non-scientific’ to many. However, rigorous mathematical analysis has proved these skeptics wrong. The simplicity of the concept has made it very popular. The idea that local interactions can generate results which are globally good is fascinating. “The concept of CA was initiated in the early 1950’s by J. Von Neumann and S. Ulam (1970). Von Neumann showed that a cellular automaton can be universal. He devised a CA, each cell of which has a state space of 29 states, and showed that the devised CA can execute any computable operation”. “A cellular automaton is a collection of “colored” cells on a grid of specified shape that evolves through a number of discrete time steps according to a set of rules based on the states of neighboring cells. The rules are then applied iteratively for as many time steps as desired. von Neumann was one of the first people to consider such a model, and incorporated a cellular model into his “universal constructor.” Cellular automata were studied in the early 1950s as a possible model for biological systems” (Wolfram 2002). Cellular Automata is a system which generates patterns that replicate themselves. The astonishing evolution of the development of these patterns forced the scientists to look at it with a whole a new perspective and not just as a game as proposed by John Horton Conway. CA generates complex pattern by simple rules and has been successfully used to reproduce any machine described in the program including a copy of itself (Wolfram, 1984). This amazing ability led a new lease of life to the theory of self-replication and the scientific fraternity has been inching closer to creating machines that can generate machines like themselves. The mechanism of working of CA is as follows. In a CA, the state of an individual cell changes with time. The state at an instance depends on the present state of that cell and that of the neighbors. These local interactions result in fascinating patterns. These states are governed by rules. Elementary CA is one dimensional automaton with two possible states per cell. These states are 1 or 0. 1 represents a live cell and 0 represents a dead cell. A cell’s neighbors defined as the cell on either side of it. A cell and its two neighbors form a neighborhood of 3 cells, so there are 23 i.e. 8 possible patterns for a neighborhood. So, there are 28 i.e. 256 possible rules. However, there are versions of CA which consider more than two neighbors also. The rules generate fascinating pattern which are not just stunning but can also form the basis of generating initial population for GAs after minor modification. The above concept has been used as the starting point of the work proposed. The number of rules depends on the number of neighbors. For instance, a three neighborhood CA would have 23 possible combinations and hence 28, that is 256 rules. The Universal nature of CA has been proved by Von Neumann. As a matter of fact, the rules given by Von Neumann were never implemented. The present popularity of CA’s can be attributed to “The Game of Life”, given by Convoy (Gardner, 1970). The formulation of the present set of rules is attributed to Stephen Wolfram (Wolfram, 1984). The detailed and involved work of Wolfram formed the basis of the practical implementations on which many complex simulations were based.

183

Applicability of Cellular Automata in Cryptanalysis

As per Biswas (Sikdar, 2000) researchers should be able to predict the global behavior from the local CA rules. Once this goal is achieved, one should be able to design the local rules/initial conditions from a given prescribed global behavior.

4. PROPOSED WORK The work proposes a model to find the key, given a sample having plaintext and cipher text, by making use of a model based on CA. Primarily, the task reduces to finding the requisite patterns. Here it may be stated that, regression using a machine learning technique can also accomplish the above task. However, the task here is to develop a mechanism for regression using CA. The work is based on the ability of Cellular automata to generate and identify patterns. The task has been accomplished as follows. First of all, the bit string for the plaintext and the cipher text are converted into patterns (note that the strings are only 0s and 1s). The crafting of these patterns would form the basis of the rest of the work. These patterns are then mapped to the most similar ones using a neural network based pattern recognizer. The step relies on the ability of a Neural Network to locate similar patterns. The appropriate rule number and the row number of the most similar pattern are then extracted. In the results section it has been stated that rule number 110 (Figure 1) proved to be one of the most useful rule for accomplishing this task (Figure 1). Figure 1. Rule 110

184

Applicability of Cellular Automata in Cryptanalysis

This is followed by preparation of a table consisting of the above data. This data forms the basis of the further analysis. To summarize the formation of patterns using the given text and the corresponding cipher text is followed by the identification of the row number and the rule number of CA. This precarious task can be carried out using Neural Networks. The model is trained using 70% of the data collected. The trained model is then verified and validated using 30% of the above data.

4.1. Analysis of the Data The collected data was a set of a 3,000 plaintext (16 bit) and corresponding cipher text (16 bits), the key (16 bit) was also used in order to train the model. The data was divided into two parts. 70% data was used for training the model and 30% for verification. The data used for training had three fields: plaintext, cipher text and the key. A connectionist, back propagation model using the CA rule and the cipher text was then developed, the key obtained and the actual key were XORed to find the Hamming distance between them. Less the Hamming distance, better the model. The coefficients of the model were assigned and modulated using the above Hamming distance. The process has been depicted in Figure 2. The model was then used to find the keys for the rest of 30% data (Figure 3). The procedure has been elucidated in Algorithm 1. The above process has been summarized in the following Algorithm (Algorithm 1) which takes a set of Plaintexts and Cipher texts as input. The process has been explained in the above discussion and depicted in Figure 2 and 3. Algorithm 1. Cellular Automata Based Cryptanalysis System Input: Plaintext (Pi), Key (Ki), Ciphertext (Ci) 1. Apply Neural Network Based Pattern similarity module to find the appropriate rule number (ri) and row number (rowi) from the repository of the CA based patterns. 2. Populate the data table from rule number (ri) and row number (rowi) and the Ki. 3. Apply NN Based Regression analyser. 4. Find the Hamming distance between the unknown keys obtained and the Kis. 5. Reconfigure the NN Feedback model as per the above results.

The above model has been implemented and results are encouraging. The results obtained have been discussed in the next section.

5. RESULTS The proposed model has been implemented and the results have been reported in this section. The discussion, conclusions and future scope have been presented in the next section.

185

Applicability of Cellular Automata in Cryptanalysis

Figure 2. CA based cryptanalysis system (CABCS): Training Phase

Figure 3. CA based cryptanalysis system (CABCS): Verification Phase

5.1. Data Collection As stated in the previous section, the data collection was done with utmost care and through analysis. The data was collected as follows. 100 English text documents of about 500 words were collected. This was followed by randomly selecting around 3000 syllables. These were then converted into ASCII values and the sum was converted into a 16-bit binary number.

5.2. Encryption The data was encrypted using a modified DES simulator developed in C#. The text obtained was of 16 bits. The binary data was then XORed with the keys to obtain the cipher text.

5.3. Model Application The Cellular Automata Based Cryptanalysis System (as explained in the previous section) was then applied to the data. The results obtained were as follows.

186

Applicability of Cellular Automata in Cryptanalysis

5.4. Results The system used pre-crafted tools for NN based pattern recognizer, and modified regression analyzer. The results were as follows (Table 2). As stated earlier 3000 sets were taken. The model, after training, produced 1372 correct keys. The number of keys that were differing in just one bit were 654. The number of keys having Hamming distance 2 were 214, and those having a Hamming distance 3 to 5 were 721. The results have been summarized in Figure 4 and Figure 5. As against this, in the model employing genetic algorithms 1201 correct keys were obtained. The number of keys differing in a single bit were 543. The comparison has been shown in Figure 6. In the figure the dark columns depict the data of the proposed work and the light colored columns depict the data of the model that uses GA. The results obtained were encouraging and pave way of the application of Neural Networks and Cellular Automata in cryptanalysis. Moreover, as compared to the model employing genetic algorithms, the results are better.

Table 2. Results Result

Number

Percentage

Number of Plaintext, keys and Ciphertexts

3000

Correct results Obtained

1372

45.73333333

Results obtained with 1 incorrect bit

654

21.8

Results obtained with 2 incorrect bit

214

7.133333333

Results obtained with 3-5 incorrect bits

721

24.03333333

Rest

39

1.3

Figure 4. The results (Pi Chart)

187

Applicability of Cellular Automata in Cryptanalysis

Figure 5. Number of results obtained in various categories

Figure 6. Comparison of the proposed technique and the technique that uses GA

6. CONCLUSION AND FUTURE SCOPE Neural networks are known for their knack to recognize patterns. This is the reason why the proposed model uses Neural Networks for pattern matching. The ability of Neural Networks to carry out regression is also well known. The work uses back propagation Neural Network for regression. The premise of the work is the ability of the CA to map patterns, which has ability to map and reproduce data. This ability has been used in various disciplines including social sciences. This makes a good case for using CA in cryptanalysis. The proposed model was developed, implemented, trained and verified. The results have been presented in the previous section and are encouraging. Though the exact keys were not obtained in

188

Applicability of Cellular Automata in Cryptanalysis

all the cases, the work at establishes the ability of CA based model to predict the keys, given sufficient data. The work is now been extended and the various types of Neural Network based models are being used both for pattern matching and regression. The extension of this work would test the model with AES keys. Moreover, the number of bits in the new setup would be increased to 128. It is also being analyzed so as to why some patterns are more helpful in cryptanalysis as compared to others. A novel model involving Diploid Genetic Algorithm (DGA) is being developed. In order to extend the work, a comprehensive literature review has already been carried out (Bhasin, 2015). The technique has also been applied on a NP Hard problem and the results are encouraging (Bhasin, 2015). The future work would work on a model which uses DGA and CA to carry out cryptanalysis. The data for verification and validation has also been collected.

REFERENCES Alani, M. M. (2012). Neuro Cryptanalysis of DES and Triple DES. Neural Information Processing, 7667, 637–646. doi:10.1007/978-3-642-34500-5_75 Bao, F. (2004). Cryptanalysis of a partially known cellular automata cryptosystem. IEEE transaction on computers, 53(11), 1493-1497. Bhasin, H., & Mehta, S. (2015). On the applicability of diploid genetic algorithms. AI & society, 31(2), 265-274. Bhasin, H., & Arora, N. (2010). Cryptography using genetic algorithms. Bhateja, A. (2014). Genetic Algorithm with elitism for cryptanalysis of vigenere cipher, Issues and challenges. Paper presented at ICICT (pp. 373-377). Gardner, M. (1970). Mathematical Games – The fantastic combinations of John Conway’s new solitaire game “life”. Scientific American. 223, 120–123. Girish, M., & Kaur, S. (2015). Cryptanalysis of transposition cipher using hill climbing and simulated annealing. Proceedings of International conference on soft computing for problem solving, Advances in intelligent systems and computing (Vol. 336, pp. 293-302). DOI:10.1007/978-81-322-2220-0_23 H. Bhasin et., al. (2015). On the applicability of diploid genetic algorithms in dynamic environments. Proceedings of the 2014 international conference on Soft computing and machine intelligence (pp. 94-97). Li, S., Li, C., Chen, G., Bourbakis, N.G., & Lo, K.-T. (2008). A general quantitative cryptanalysis of permutation-only multimedia ciphers against plaintext attacks. Signal Processing Image Communication, 23(3), 212-223. DOI:1016/j.image.2008.01.003 Lin, F.-T., & Kao, C.-Y. (1995). A genetic algorithm for cipher text only attack in cryptanalysis, Systems, man and cybernetics. In Intelligent systems for the 21st century (Vol. 1, pp 650-654). DOI:. doi:10.1109/ ICSMC.1995.537837

189

Applicability of Cellular Automata in Cryptanalysis

Luthra J. (2011, December). A Hybrid firefly algorithm using genetic operators for the cryptanalysis of a mono alphabetic substitution cipher. Proceedings of the Information and communication technologies (pp. 202-206). Doi:10.1109/WICT.2011.6141244 M. din et., al. (2014). Cryptanalysis of Geffe Generator Using Genetic Algorithm. Proceedings of the third international conference on soft computing for problem solving, advances in intelligent systems and computing (pp. 509-514). Doi:10.1007/978-81-322-1768-8_45 Manjula, R., & Anitha, R. (2011). Identification of Encryption Algorithm using Decision tree. Advanced computing communications in computer and information science, 133, 237-246. Doi:10.1007/978-3642-17881-8_23 Omran, S., Al-Khalid, A.S., & Alsaady, D. (2010, December). Using Genetic Algorithm to Break a Mono-Alphabetic Substitution Cipher. Proceedings of ICOS IEEE Conference (pp. 63-67). DOI:10.1109/ ICOS.2010.5720065 Li, C., Li, S., Alvarez, G., Chen, G., & Lo, K.-T. (2007). Cryptanalysis of two chaotic encryption schemes based on circular bit shift and XOR operations. Physics Letters. [Part A], 369(1-2), 23–30. doi:10.1016/j. physleta.2007.04.023 Rothe, J. (2002). Some facets of complexity theory and cryptography: A five-lecture tutorial. ACM Comput. Surv. 34(4), 504-549. Doi:10.1145/592642.592646 Schneier, B. (1995). Applied cryptography (2nd ed.). New York: John Wiley & Sons, Inc. Seredynski, F., Bouvry, P., & Zomaya, A. Y. (2004). Cellular automata computation and secret key cryptography. Parallel Computing, 30(5-6), 753–766. doi:10.1016/j.parco.2003.12.014 Sikdar, B. K., Paul, K., Biswas, G. P., Yang, C., Boppana, V., Mukherjee, S., & Pal Chaudhuri, P. (2000, January). Theory and 28 Application of GF(2p) Cellular Automata as On-Chip Test Pattern Generator. Proceedings Intl. Conf. on VLSI Design, India (pp. 556–561). Song, J., Zhang, H., Meng, Q., & Wang, Z. (2007). Cryptanalysis of four-round DES Based on Genetic algorithm. Wireless communications. Networking and Mobile Computing. Doi:10.1109/WICOM.2007.580 Sung, J., Hong, D., & Hong, S. (2007). Cryptanalysis of an involutional block cipher using cellular automata. Information processing letters, 104(5), 183-185. Tumassini, M., & Perrenoud, M. (2001). Cryptography with cellular automata. Applied soft computing, 1(2), 151-160. Uddin, M.F., & Youseef, A.M. (2006). Cryptanalysis of simple substitution ciphers using particle swarm optimization. Evolutionary computation (pp. 677-680). Doi:10.1109/CEC.2006.1688376 Ulam, S. (1970). Some mathematical problems connected with patterns of growth figures. In A. W. Burks (Ed.), Essays on Cellular Automata. Illinois: Univ. Illinois Press.

190

Applicability of Cellular Automata in Cryptanalysis

Wang, X., Wang, Q., & Zhang, Y. (2015). A fast image algorithm based on rows and columns switch. Nonlinear Dynamics, 79(2), 1141-1149. Doi:10.100711071-014-1729-y Wolfram, S. (1984). Computation theory of cellular automata. Communications in Mathematical Physics, 96(1), 15–57. doi:10.1007/BF01217347 Wolfram, S. (2002). A new kind of science. Champaign, IL, USA: Wolfram Media.

This research was previously published in the International Journal of Applied Metaheuristic Computing (IJAMC), 8(2); edited by Peng-Yeng Yin; pages 38-48, copyright year 2017 by IGI Publishing (an imprint of IGI Global).

191

Section 3

Encryption Keys and Homomorphic Encryption

193

Chapter 12

A Novel Approach of Symmetric Key Cryptography using Genetic Algorithm Implemented on GPGPU Srinivasa K. G. M. S. Ramaiah Institute of Technology, India Siddesh G. M. M. S. Ramaiah Institute of Technology, India Srinidhi Hiriyannaiah M. S. Ramaiah Institute of Technology, India Anusha Morappanavar M. S. Ramaiah Institute of Technology, India Anurag Banerjee M. S. Ramaiah Institute of Technology, India

ABSTRACT The world of digital communication consists of various applications which uses internet as the backbone for communication. These applications consist of data related to the users of the application, which is confidential and integrity needs to be maintained to protect against unauthorized access and use. In the information hiding field of research, Cryptography is one of the wide techniques used to provide security to the internet applications that overcome the challenges like confidentiality, integrity, authentication services etc. In this paper, we present a novel approach on symmetric key cryptography technique using genetic algorithm that is implemented on CUDA architecture.

DOI: 10.4018/978-1-7998-1763-5.ch012

Copyright © 2020, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

A Novel Approach of Symmetric Key Cryptography using Genetic Algorithm Implemented on GPGPU

INTRODUCTION In the internet era of applications, confidentiality, security, integrity and authentication services are increasingly becoming more important (Viega & McGraw, 2001). One of the key techniques used for providing secure communication is cryptography. Cryptography generally deals with exchange of messages between the sender and the receiver using some secure keys with encryption (encoding) and decryption (decoding). A brief introduction of cryptography, key components of it and its different types are discussed in section 1. In the evolution of nature, a biological entity that adapts to the changes in the environment has better chances of survival according to the Darwin’s theory of evolution. This analogy is applied to genetic programming, which uses theory of evolution steps to draw a better solution to a problem being solved by an algorithm (Kahn, 1996). The steps involved in genetic algorithm or programming is discussed in section 2. With the advent of increasing multi-core processors, applications being developed need to utilize the threads functionality of these processors. Compute Unified Device Architecture (CUDA), a unified programming model was introduced by Nvidia, which facilitates programming both sequential and parallel portions of a program within a single unit (Nickolls et al., 2008). The different components of CUDA and its architecture are discussed in section 3. Random numbers are generally used in cryptography for encryption and decryption. Many methods can be used in generating random numbers. The method of creating pseudo- random numbers using genetic algorithms involves more computation power, which can be processed using GPU using CUDA architecture. These pseudo random numbers are generated with linear congruential method. With the help of genetic algorithms implemented using CUDA architecture, the pseudo random numbers are used for encryption and decryption. The algorithm is compared with computation time spent on the CPU and the results are encouraging with CUDA. The paper is organized as follows. In section 1 we discuss briefly concepts related to cryptography, section 2 on genetic algorithms, section 3 on CUDA, section 4 on pseudo random generation and in the later sections, proposed approach and experimental results are discussed.

1. CRYPTOGRAPHY In our daily life of internet applications and email systems, keeping data and messages confidentially is more important, for example in the launch of nuclear codes and other mission critical systems, in spy’s profession the data confidentiality is not compromised. The science of protection of data and communications is called Cryptography (Viega & McGraw, 2001) and (Stinson, 2005). There are many applications where cryptography is applied currently in the fields of e-commerce transactions where examples include purchase using credit cards, wire money transfer etc. In this section, we discuss basic terminologies that are used in cryptography. The transmission of a message involves two key elements namely the sender and the receiver. In Cryptography the messages are sent between sender and the receiver using encryption and decryption techniques, ensuring the information is received by the intended receivers without any intruders in the middle of the communication (Kessler, 2015). The basic model and different terminologies used in the cryptography are as shown in the Figure 1. 194

A Novel Approach of Symmetric Key Cryptography using Genetic Algorithm Implemented on GPGPU

• • • •

Plain Text: The message that needs to be transmitted from the sender to the receiver in the original form. Encryption: A phase where the original message is encoded using a key. Cipher Text: The text that is the result of the encryption phase. Decryption: A phase of extracting the original message from the encoded message using a key.

The two basic methods of cryptography are symmetric and asymmetric cryptography (Stanoyevitch, 2010) and (Hellman, 2002). In symmetric cryptography, both sender and the receiver share the same key for encryption and decryption respectively as shown in the Figure 2 (Hellman, 2002). Thus, the key shared in this method needs to be more secure and powerful. In asymmetric cryptography, both sender and the receiver share different public and private keys for encryption and decryption respectively. Cipher text plays an important role in cryptography, which must be kept secret and secure from the intruders. The analysis and study of breaking cipher text is called cryptanalyis (Barak et al., 2001). In order to generate keys required for the encryption and decryption, random numbers can be generated and same can be used as keys. These random numbers should be large in number and feasible enough to keep away from the intruders (Koblitz, 1994). Genetic programming can be used in evolving the population of random numbers using different techniques of it (Barker & Kelsey, 2012). In our approach, the random numbers are generated using linear congruential method and population is evolved using genetic algorithms. We discuss some of the basic steps followed in developing a genetic algorithm in the next section. Figure 1. Basic model of cryptography

Figure 2. Symmetric key cryptography

195

A Novel Approach of Symmetric Key Cryptography using Genetic Algorithm Implemented on GPGPU

2. GENETIC ALGORITHMS In nature, in the evolution of life, adaptation is always a key survival entity related to any biological life. Darwin’s theory of evolution is applied for the survival of the fittest in the environment where biological entities exist. This theory of evolution can be applied to solving computer problems called as Genetic programming (Mitchell, 1999). The degree of adaptation of a solution to a given problem determines the fitness of the solution for a given population of inputs to the problem. A new population of solution might emerge that depends on the degree of adaptation determined (Thede, 2004). In this section, we describe in brief about genetic programming and its steps. A genetic algorithm is a type of searching algorithm that uses the concept of survival of the fittest in the theory of evolution [10]. It searches a solution for a given problem using a list of the populations of solutions. It starts initially with a population of solutions, then evolves this population of solution using techniques of genetic evolution (cross over & mutation) in (Thede, 2004) and (Koza, 1992). The algorithm is terminated with a suitable condition when the best solution of population is achieved. The generic steps involved in the genetic algorithm are as follows and is as shown in the Figure 3. 1. START: Generate random population of n individuals (using some suitable method). 2. FITNESS: Evaluate the fitness f(x) of each individual x in the population. 3. NEW POPULATION: Create a new population by repeating following steps until the new population is complete. a. SELECTION: Select two parent individuals from a population according to their fitness (the better fitness, the bigger chance to be selected). b. CROSSOVER: With a crossover probability, cross over the parents to form a new offspring. c. MUTATION: With a mutation probability, mutate the offspring at some locus (position in chromosome). d. ACCEPTING: Place new offspring in a new population. 4. REPLACE: Use new generated population for a further run of algorithm 5. TEST: If the end condition is satisfied, stop and return the best solution in the current population. 6. LOOP: Go to step 2. The population of solutions is the initial collection of solutions considered which can be generated using random methods like linear congruential method, random number streams etc. An individual solution is a single solution in the population. The fitness of an individual solution produces the absolute value of the solution, which evaluates the fitness priority of the individual over other individual solutions in the population (Mitchell, 1999). Once the fitness is evaluated, the algorithm continues to produce a new set of population using selection, crossover and mutation methods (Thede, 2004) and (Koza, 1992). Crossover is achieved by selecting a random locus on the pair of parents and their bits are exchanged to generate two new offspring. The most used way of encoding is a binary string which is as shown using examples of parent 1 and parent 2. Parent 1 = 1101100100110110 Parent 2 = 1101111000011110

196

A Novel Approach of Symmetric Key Cryptography using Genetic Algorithm Implemented on GPGPU

Figure 3. Steps in genetic algorithm

One of the ways to do crossover is to choose some crossover point randomly. Everything before this point is copied from first parent and everything after this point is copied from second parent [10] [11]. An example is as shown below. Example: If the crossover point is chosen as the position after 5th bit, then crossover looks like: Parent 1 Parent 2 Offspring Offspring

1 2

11011 | 00100110110 11011 | 11000011110 11011 | 11000011110 11011 | 00100110110

197

A Novel Approach of Symmetric Key Cryptography using Genetic Algorithm Implemented on GPGPU

Box 1.

This is known as Single point crossover (Koza, 1992)– one crossover point is selected, binary string from the beginning of the chromosome to the crossover point is copied from one parent and the binary string from the crossover point to the end is copied from another parent (Box 1). Mutation is achieved by selecting a random bit in the offspring and inverting it. The new offspring generated are placed in the next generation (Mitchell, 1999), (Thede, 2004) and (Koza, 1992). After a crossover is performed, mutation takes place. Mutation changes the new offspring randomly. For binary encoding, we can switch a few randomly chosen bits from 1 to 0 or from 0 to 1 which as shown in the example below. Example: If bit 5 is chosen randomly then, mutation would be as follows: Offspring

1100011100000111

After mutation, Offspring

1100111100000111

Genetic algorithms require faster computation time, as it involves generating initial population, apply cross over and mutation techniques on it. In this regard, we can use CUDA a GPU programming model that facilitates in carrying out large computations with more number of threads. We discuss the CUDA programming model in the next section.

3. CUDA With the advent of the multi core processors and many core GPUs and their increasing speed, applications needs to be developed that are scalable enough on these processors. In 2007, nvidia released CUDA,

198

A Novel Approach of Symmetric Key Cryptography using Genetic Algorithm Implemented on GPGPU

a scalable programming model that allows computation to be carried out both on CPU and many core GPUs (NVIDIA, 2007). Many applications related to computation, searching, bio-medical imaging sparse matrices etc are being rapidly developed using the CUDA programming model (Luebke, 2008). In this section, we briefly discuss some of the related key concepts to CUDA. The three basic features that are offered by CUDA are hierarchy of thread groups, shared memories and barrier synchronization (Nickolls et al., 2008). In developing a parallel algorithm for a problem using CUDA, first the portion of the problem that can be done independently needs to be identified, and further execute these parts across independent parallel threads (Kirk & Hwu, 2012). CUDA is a minimal extension to C and C++ where the functions or full program are kernels. The kernels are executed across parallel threads that are organized using thread blocks and grids. A thread block is a set of threads that executes with barrier synchronization and a shared private memory space among themselves (Nickolls et al., 2008) and (Kirk & Hwu, 2012). A grid is a set of thread blocks that executes independently and parallel. A basic organization of the CUDA with threads, blocks and grids is as shown in the Figure 4. The programmer has to specify the number of threads per block and the number of blocks in the grid. CUDA supports 512 threads in one thread block. Each thread is given a unique id threadIdx numbered from 0,1…..blockdim-1 within a thread block. Each block is given a unique id blockIdx in the grid. In order to execute the kernels on CUDA, the data that resides on the host needs to be transferred to the device or GPU first, invoke the kernel with appropriate number of threads and then finally execute it on the device (Kirk & Hwu, 2012). To perform these operations certain keywords are used that suggest the kernels are executed on the device and not on the host. Some of the keywords used are cudaMemcopy(), _global_, .We discuss some of them below that have been used in carrying out in our approach. A CUDA kernel is recognized by the keyword __global. The specifications regarding the number of blocks in the grids and the number of threads in the blocks are provided in the kernel call. To compute in parallel, a number of threads need to be generated while invoking the kernel. CUDA uses dimGrid and dimBlock as the keywords for the parameters in the kernel function to specify the number of threads (NVIDIA, 2015). The kernel call is of the form Kernel(parameters..)

dimGrid indicates the number of blocks to be invoked in a grid while dimBlock indicates the number of threads to be invoked in a block. More the number of threads, more computation are done in parallel. Once dimGrid and dimBlock are fixed, the threads have to be assigned with their index value to ensure that each thread handles its own data. CUDA provides some keywords like threadIdx, blockIdx and blockDim (NVIDIA, 2015). For a one-dimensional array, the index can be assigned as, int tid = threadIdx.x + blockIdx.x * blockDim.x

These indices can be incremented as, tid += blockDim.x * gridDim.x;

199

A Novel Approach of Symmetric Key Cryptography using Genetic Algorithm Implemented on GPGPU

Figure 4. CUDA thread blocks and grids

This makes sure that all the threads compute on their respective data without interfering with other data. As these threads work simultaneously, the entire computation can be done much faster than its sequential counterpart. The kernel gets its input data from the host and the result must be sent back to the host. To accomplish this, data needs to be transferred between the host and the device. cudaMemcpy() allows the data transfer between host and the device (NVIDIA, 2015). cudaMemcpy(dest,src,size,type_of_transfer)

200

A Novel Approach of Symmetric Key Cryptography using Genetic Algorithm Implemented on GPGPU

It copies “size” bytes from the memory area pointed to by “src” to the memory area pointed to by “dest”, where “type_of_transfer” specifies the direction of transfer. The CPU and GPU do not have shared memory. They compute on variables that are allocated in their own memory. In CPU, memory can be allotted to variables by simple declaration or by using C function malloc(). The variables to be used in the kernel function must be allocated memory in the host code before invoking the kernel function. This is done by using the function cudaMalloc(). We may want to transfer data from host variables to device variables so that they can be sent as parameters to the kernel function. To do this, cudaMemcpy() is used. Once memory is allotted and data transfer is complete, the kernel is invoked by calling the kernel function. In this function call, number of blocks and number of threads should be specified. Parameters can also be sent. The variables allocated using cudaMalloc() must be sent as pointers. After the execution of kernel function, the data from device variables must be transferred to host variables for further processing. This is again done using cudaMemcpy() function. The maximum overhead incurred in a GPU code is during the data transfer between host and device. To overcome this overhead, CUDA provides streams to hide the latency(Kirk & Hwu, 2012). The data to be transferred is allocated in the host pinned memory. Copies between pinned memory and device memory can be performed concurrently with kernel execution. Pinned memory can be mapped into the address space of the device, eliminating the need to copy. Allocation in pinned memory is done using cudaHostAlloc() function. Simultaneous data transfer can be done by creating CUDA streams. Instructions in the same stream cannot be executed simultaneously. Instructions in different streams can overlap to hide the data transfer latency. cudaMemcpyAsync() function is used to carry out simultaneous data transfer (NVIDIA, 2015). In our approach on the parallel implementation of the encryption and decryption using genetic algorithms CUDA model is used for faster computation which is discussed in section 5 and 6.

4. LINEAR CONGRUENTIAL METHOD OF GENERATING PSEUDO RANDOM NUMBERS A cryptographic algorithm requires a series of numbers to be produced for encryption and decryption phases. Random number generators are used to produce a stream of numbers for cryptographic algorithms [20]. The random numbers generated usually have two properties namely, uniformity and independence i.e. the random numbers generated are independent from the given continuous sample range. The random numbers generated for computer algorithms are called as pseudo-random numbers that deceive the user; the random numbers does not appear random. One of methods used to generate pseudo-random numbers is linear congruential method (Barker & Kelsey, 2012). The Linear Congruential Method requires four parameters namely Xn, a, c and m. The random numbers are generated by using the following equation iteratively: Xn+1 = (aXn+c) mod m Therefore, generation 1 will be Xo = Xn

201

A Novel Approach of Symmetric Key Cryptography using Genetic Algorithm Implemented on GPGPU

Xn+1 = (aXo+c) mod m etc. where Xo is the initial seed, ‘a’ is constant multiplier, c is the increment and m is modulus.

5. PROPOSED APPROACH The project consists of two parts: encryption and decryption on the sender and receiver side respectively. Both encryption and decryption require the generation of the array of pseudorandom numbers. We make use of a structure to store the binary representation of the chromosomes and their respective fitness values. typedef struct Chrom { int bit[64]; int fit; int value; }chrom;

The sender and receiver exchange 8 parameters through a secure link. These are: • • •

Xn, a, c, and m: These are used in Linear Congruential Method. Start_number and end_number: These are used to obtain a subset of the pseudorandom numbers Remainder and Modulus: These are used to enforce another level of security.

Pseudo random numbers are generated using linear congruential method- sequential. The sequence of random numbers are dependent on one another, hence it is sequentially programmed. These random numbers are selected by evaluating their fitness using GA with steps on cross over and mutation, and done on CUDA architecture. Algorithms 1 to 5 discusses different operations that are followed in genetic algorithms, are presented below which are represented as kernels in CUDA program. •

Fitness: Once the array of pseudorandom numbers is created, they are stored in the structure in their binary representation. Their fitness value is calculated by checking if the number is a prime, if it is a prime, a fitness value of 1 is assigned to its structure. If not, a fitness value of 0 is assigned to it. The algorithm for fitness is shown in Algorithm 1.

Algorithm 1. Fitness function Input: array of N numbers 1. for i=0 to N 2. flag=0 3. for j=2 to array[i] 4. if (array[i] % j = 0) 5. flag = 1 6. if (flag = 0)

202

A Novel Approach of Symmetric Key Cryptography using Genetic Algorithm Implemented on GPGPU

7. chromosome[i].fit = 1 8. else 9. chromosome[i].fit = 0

Once the first generation is created, it undergoes genetic algorithm operations to create the next generation of new pseudorandom numbers. The first generation numbers are used in three operations namely survival of the fittest, crossover and mutation. •

Survival of the Fittest: The first generation is passed as an argument to the kernel function and the pseudorandom numbers are selected based on their fitness value. The algorithm for the survival is shown in Algorithm 2 (Delman, 2004).

Algorithm 2. Survival of the fittest Input: Array of N numbers with fitness value 1. size=0 2. for i=0 to N 3. if (chromosome[i].fit = 1) 4. array[size] = chromosome[i] 5. size = size + 1

Thus, the chromosomes of current population are selected according to their fitness value. •

Crossover: The selected generation undergoes crossover (Algorithm 3). The simplest way to do this is to choose some crossover point randomly. Everything before this point is copied from the first parent and everything after the crossover point is copied from the second parent (Bergmann et al., 2008).

Algorithm 3. Crossover Input: Array of selected numbers 1. k = 0 2. Choose a random point in the chromosome 3. for j=0 to size 4. for i =0 to random 5. next[k].bit[i]=current[k].bit[i] 6. next[k+1].bit[i]=current[k+1].bit[i]; 7. for i = random to 63 8. next[k].bit[i]=current[k+1].bit[i]; 9. next[k+1].bit[i]=current[k].bit[i]; 10. calculate fitness

Thus, the result of crossover is stored in next population.

203

A Novel Approach of Symmetric Key Cryptography using Genetic Algorithm Implemented on GPGPU

•

Mutation: This generation of numbers undergoes mutation with at low probability of 2%. Single point mutation occurs i.e. a random point is chosen and that bit is inverted (Algorithm 4).

Algorithm 4. Mutation Input: New population 1. Choose a random point 2. row = random%size 3. col = random%66 4. if (next[row].bit[col] = 1) 5. next[row].bit[col] = 0 6. else 7. next[row].bit[col] = 1 8. calculate fitness

Thus, the result of mutation is stored in the new population. Finally, the current population and next population are put together to form an array of pseudorandom numbers. To form the key, a subset of this array is selected depending on the start_number and end_number given by the sender and receiver (Algorithm 5). To enforce another level of security, only those numbers which leave remainder remainder when divided by modulus are selected (Dorrendorf et al., 2007). Algorithm 5. Creating sub-array Input: Array of pseudorandom numbers 1. for i=0 to array_size 2. if array[i]>=start_number 3. if array[i] 0. Definition 4: (Contrast Condition for share image) The share image S reveals the content of the cover image R if α > 0. Definition 5: (Security Condition, Chen and Tsao (2011)) A (K,n) RG-based Visual Cryptography Scheme is secure if the stacking of any r (r